[Met_help] [rt.rap.ucar.edu #71078] History for MET Point-Stat mask_grid issue
John Halley Gotway via RT
met_help at ucar.edu
Fri Mar 13 15:53:32 MDT 2015
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
I'm using MET V4.1 and tried to specify use of a forecast grid (GRIB1 format) to define a verification masking region per the guidance on the User's Guide on p. 4-12 ("...any gridded data file that MET can read..." . I used the following entry in the config file:
mask = {
grid = ["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02-07_12:00:00"];
poly = [];
sid = [];
};
The ERROR dialog is as follows:
DEBUG 4: parse_grid_mask() -> parsing grid mask "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02-07_12:00:00"
ERROR :
ERROR : parse_grid_mask() -> the mask_grid requested "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02-07_12:00:00" is not defined.
ERROR :
The User's Guide mentions that the user must specify a description of the field to be used from the input file, but I wasn't sure how that info is input.
Do you have any suggestions I can try to resolve this issue?
Thanks.
John Raby
US Army Research Laboratory
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue UPDATED info
From: Raby, John W USA CIV
Time: Thu Mar 12 13:17:23 2015
I was able to locate some info in the README file in the software
distribution folder which describes the entries in the config file to
describe the field in the forecast file to be used. I modified my
config file as follows:
mask = {
grid = [];
poly =
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_12:00:00 {name = \"TMP\"; level = \"Z2\";}"
];
sid = [];
};
Point-Stat executed for a longer period, but no output. Had an ERROR
as follows:
ERROR :
ERROR : parse_poly_mask() -> the masking grid does not match the
input grid.
ERROR :
This appears to present an issue with the way I want to use the
mask_grid. My intent was to score the input forecast file, but only
over a smaller domain centered inside the domain of the input forecast
file. Any thoughts on this development and how I can proceed?
Thanks.
John
_______________________________________
From: met_help at ucar.edu via RT [met_help at ucar.edu]
Sent: Thursday, March 12, 2015 11:44 AM
To: Raby, John W CIV USARMY ARL (US)
Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat mask_grid
issue
Greetings,
This message has been automatically generated in response to the
creation of a trouble ticket regarding:
"MET Point-Stat mask_grid issue",
a summary of which appears below.
There is no need to reply to this message right now. Your ticket has
been
assigned an ID of [rt.rap.ucar.edu #71078].
Please include the string:
[rt.rap.ucar.edu #71078]
in the subject line of all future correspondence about this issue. To
do so,
you may reply to this message.
Thank you,
met_help at ucar.edu
-------------------------------------------------------------------------
I'm using MET V4.1 and tried to specify use of a forecast grid (GRIB1
format) to define a verification masking region per the guidance on
the User's Guide on p. 4-12 ("...any gridded data file that MET can
read..." . I used the following entry in the config file:
mask = {
grid =
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_12:00:00"];
poly = [];
sid = [];
};
The ERROR dialog is as follows:
DEBUG 4: parse_grid_mask() -> parsing grid mask
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_12:00:00"
ERROR :
ERROR : parse_grid_mask() -> the mask_grid requested
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_12:00:00" is not defined.
ERROR :
The User's Guide mentions that the user must specify a description of
the field to be used from the input file, but I wasn't sure how that
info is input.
Do you have any suggestions I can try to resolve this issue?
Thanks.
John Raby
US Army Research Laboratory
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: Raby, John W USA CIV
Time: Fri Mar 13 08:02:18 2015
Classification: UNCLASSIFIED
Caveats: NONE
I tried another method to define the masking region using the
gen_poly_mask
tool using my input forecast file and a poly file which defines the
rectangular subdomain in the center of the two larger input forecast
domains
and was successful, but the results I get are no different than when I
use
the poly text file itself. This brings up another question about the
results
I get with the poly text file. If I run Point-Stat with three
different
input forecast grids (all generated from the same WRF run over a set
of
nested domains which have a common center) with three different
resolutions
(grid spacing) and specify the same masking region using the text poly
file
which defines the domain of the innermost nest, the three CNT output
files
show three different values (621, 619 and 617) for the TOTAL number of
matched pairs for the variable TMP/Z2. Why is this the case when the
scoring
is being done over the same masking region? My expectation was that
the
TOTAL number of matched pairs should all be the same.
Thanks.
R/
John
-----Original Message-----
From: Raby, John W CIV USARMY ARL (US)
Sent: Thursday, March 12, 2015 1:17 PM
To: met_help at ucar.edu
Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
UPDATED
info
I was able to locate some info in the README file in the software
distribution folder which describes the entries in the config file to
describe the field in the forecast file to be used. I modified my
config
file as follows:
mask = {
grid = [];
poly =
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_1
2:00:00 {name = \"TMP\"; level = \"Z2\";}"
];
sid = [];
};
Point-Stat executed for a longer period, but no output. Had an ERROR
as
follows:
ERROR :
ERROR : parse_poly_mask() -> the masking grid does not match the
input
grid.
ERROR :
This appears to present an issue with the way I want to use the
mask_grid.
My intent was to score the input forecast file, but only over a
smaller
domain centered inside the domain of the input forecast file. Any
thoughts
on this development and how I can proceed?
Thanks.
John
_______________________________________
From: met_help at ucar.edu via RT [met_help at ucar.edu]
Sent: Thursday, March 12, 2015 11:44 AM
To: Raby, John W CIV USARMY ARL (US)
Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat mask_grid
issue
Greetings,
This message has been automatically generated in response to the
creation of
a trouble ticket regarding:
"MET Point-Stat mask_grid issue", a summary of which appears
below.
There is no need to reply to this message right now. Your ticket has
been
assigned an ID of [rt.rap.ucar.edu #71078].
Please include the string:
[rt.rap.ucar.edu #71078]
in the subject line of all future correspondence about this issue. To
do so,
you may reply to this message.
Thank you,
met_help at ucar.edu
-------------------------------------------------------------------------
I'm using MET V4.1 and tried to specify use of a forecast grid (GRIB1
format) to define a verification masking region per the guidance on
the
User's Guide on p. 4-12 ("...any gridded data file that MET can
read..." . I
used the following entry in the config file:
mask = {
grid =
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_1
2:00:00"];
poly = [];
sid = [];
};
The ERROR dialog is as follows:
DEBUG 4: parse_grid_mask() -> parsing grid mask
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_12
:00:00"
ERROR :
ERROR : parse_grid_mask() -> the mask_grid requested
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_12
:00:00" is not defined.
ERROR :
The User's Guide mentions that the user must specify a description of
the
field to be used from the input file, but I wasn't sure how that info
is
input.
Do you have any suggestions I can try to resolve this issue?
Thanks.
John Raby
US Army Research Laboratory
Classification: UNCLASSIFIED
Caveats: NONE
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: John Halley Gotway
Time: Fri Mar 13 10:29:46 2015
John,
Glad you were able to make some progress. I see a couple of remaining
questions in there...
"Using the output of gen_poly_mask and the lat/lon polyline itself
produces
the same result. Is that OK?"
My answer is... that's great, and I'm relieved they do! The main
reason to
use gen_poly_mask is to make the MET statistics tools (i.e. point_stat
or
grid_stat) run a bit faster. It can be slow to figure out which
points are
inside/outside a polyline region if the polyline contains thousands of
point. In that case, you'd run gen_poly_mask to create the mask once
rather than having to recreate it each time you run a MET statistics
tool.
But if your polyline only contains a small number of points, running
gen_poly_mask first would only save you a second or two.
Next, you were trying to use a smaller grid to create a mask for a
larger
grid. The gen_poly_mask tool does support this. By the way, since
METv5.0, we've realized that gen-"poly"-mask is poorly named, and in
the
next release it'll be replaced by a more generalized tool named
gen_vx_mask. It include some more flexible options.
But if you read the usage statement for gen_poly_mask, the...
"mask_file" is an ASCII Lat/Lon polyline file or gridded data file
defining the masking region
So you can pass a gridded data file for the nest as the "mask_file"
rather
than having to create a set of lat/lon points. I'd suggest going that
route.
Here's an example of that using data from the MET tarball:
cd met-5.0
bin/gen_poly_mask \
data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
data/sample_fcst/2009123112/arw-fer-gep1/d01_2009123112_02400.grib \
mask.nc -name MY_MASK
Unfortunately, I don't think I fully understand your last question as
to
why you're getting slightly different matched pair counts (621, 619
and
617). But my guess is one of two possibilities. Either...
(1) The polyline mask for the different grids is producing slightly
different results and the differences lie along the boundary of the
mask.
(2) There's some missing data values somewhere in the forecast and
observations causing slightly different matched pairs.
I'd suggest configuring point_stat to dump out the MPR lines for those
three runs. Then we could take a close look at them to see where the
differences lie. Identifying the stations where the differences occur
is
the first step in finding an explanation.
Thanks,
John
On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
> I tried another method to define the masking region using the
gen_poly_mask
> tool using my input forecast file and a poly file which defines the
> rectangular subdomain in the center of the two larger input forecast
> domains
> and was successful, but the results I get are no different than when
I use
> the poly text file itself. This brings up another question about the
> results
> I get with the poly text file. If I run Point-Stat with three
different
> input forecast grids (all generated from the same WRF run over a set
of
> nested domains which have a common center) with three different
resolutions
> (grid spacing) and specify the same masking region using the text
poly file
> which defines the domain of the innermost nest, the three CNT output
files
> show three different values (621, 619 and 617) for the TOTAL number
of
> matched pairs for the variable TMP/Z2. Why is this the case when the
> scoring
> is being done over the same masking region? My expectation was that
the
> TOTAL number of matched pairs should all be the same.
>
> Thanks.
>
> R/
> John
>
> -----Original Message-----
> From: Raby, John W CIV USARMY ARL (US)
> Sent: Thursday, March 12, 2015 1:17 PM
> To: met_help at ucar.edu
> Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
UPDATED
> info
>
> I was able to locate some info in the README file in the software
> distribution folder which describes the entries in the config file
to
> describe the field in the forecast file to be used. I modified my
config
> file as follows:
>
> mask = {
> grid = [];
> poly =
>
>
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_1
> 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> ];
> sid = [];
> };
>
> Point-Stat executed for a longer period, but no output. Had an ERROR
as
> follows:
>
> ERROR :
> ERROR : parse_poly_mask() -> the masking grid does not match the
input
> grid.
> ERROR :
>
> This appears to present an issue with the way I want to use the
mask_grid.
> My intent was to score the input forecast file, but only over a
smaller
> domain centered inside the domain of the input forecast file. Any
thoughts
> on this development and how I can proceed?
>
> Thanks.
> John
> _______________________________________
> From: met_help at ucar.edu via RT [met_help at ucar.edu]
> Sent: Thursday, March 12, 2015 11:44 AM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
mask_grid
> issue
>
> Greetings,
>
> This message has been automatically generated in response to the
creation
> of
> a trouble ticket regarding:
> "MET Point-Stat mask_grid issue", a summary of which appears
below.
>
> There is no need to reply to this message right now. Your ticket
has been
> assigned an ID of [rt.rap.ucar.edu #71078].
>
> Please include the string:
>
> [rt.rap.ucar.edu #71078]
>
> in the subject line of all future correspondence about this issue.
To do
> so,
> you may reply to this message.
>
> Thank you,
> met_help at ucar.edu
>
>
-------------------------------------------------------------------------
> I'm using MET V4.1 and tried to specify use of a forecast grid
(GRIB1
> format) to define a verification masking region per the guidance on
the
> User's Guide on p. 4-12 ("...any gridded data file that MET can
read..." .
> I
> used the following entry in the config file:
>
> mask = {
> grid =
>
>
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_1
> 2:00:00"];
> poly = [];
> sid = [];
> };
>
> The ERROR dialog is as follows:
>
> DEBUG 4: parse_grid_mask() -> parsing grid mask
>
> "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_12
> :00:00"
> ERROR :
> ERROR : parse_grid_mask() -> the mask_grid requested
>
> "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02-07_12
> :00:00" is not defined.
> ERROR :
>
> The User's Guide mentions that the user must specify a description
of the
> field to be used from the input file, but I wasn't sure how that
info is
> input.
>
> Do you have any suggestions I can try to resolve this issue?
>
> Thanks.
>
> John Raby
> US Army Research Laboratory
>
>
>
>
>
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
>
>
>
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: Raby, John W USA CIV
Time: Fri Mar 13 11:12:47 2015
Classification: UNCLASSIFIED
Caveats: NONE
John -
Thanks. Your answers address all my questions very well.
My lat/lon polyline has only four points so I concur that the runtime
is a
non-issue for sure.
I am very interested in your suggestion about using the gen_poly_mask
tool and
providing the GRIB file of the smaller forecast as the input
"mask_file". I
just generated the usage in an xterm and see the wording you show
below. I
should have checked there. Maybe an edit change for the User's Guide
p. 3-24
(V4.1 or later update) would also help.
Regarding the different matched pair counts, I have the MPR files for
the 3
runs and could do a "diff" among them and some other exploring to look
for the
differences. I will let you know what differences I find. I suspect
that the
first possibility you mention below is the more likely explanation. If
the
input forecast grid is referenced when converting the polygon points
into the
x/y values and the same with the observation locations, then that x/y
grid
space might differ some from that which would be referenced when a
forecast
grid with closer grid spacing is used with the same polygon points and
observations.
I am hoping that by using the gen_poly_mask tool and providing the
GRIB file
of the smaller forecast as the input "mask_file", a more accurate
replication
of the desired masking region might result which results in more
consistency
in the numbers of matched pairs. I will let you know the outcome.
R/
John
-----Original Message-----
From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
Sent: Friday, March 13, 2015 10:30 AM
To: Raby, John W CIV USARMY ARL (US)
Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more
UPDATED
info (UNCLASSIFIED)
John,
Glad you were able to make some progress. I see a couple of remaining
questions in there...
"Using the output of gen_poly_mask and the lat/lon polyline itself
produces
the same result. Is that OK?"
My answer is... that's great, and I'm relieved they do! The main
reason to
use gen_poly_mask is to make the MET statistics tools (i.e. point_stat
or
grid_stat) run a bit faster. It can be slow to figure out which
points are
inside/outside a polyline region if the polyline contains thousands of
point.
In that case, you'd run gen_poly_mask to create the mask once rather
than
having to recreate it each time you run a MET statistics tool.
But if your polyline only contains a small number of points, running
gen_poly_mask first would only save you a second or two.
Next, you were trying to use a smaller grid to create a mask for a
larger
grid. The gen_poly_mask tool does support this. By the way, since
METv5.0,
we've realized that gen-"poly"-mask is poorly named, and in the next
release
it'll be replaced by a more generalized tool named gen_vx_mask. It
include
some more flexible options.
But if you read the usage statement for gen_poly_mask, the...
"mask_file" is an ASCII Lat/Lon polyline file or gridded data file
defining
the masking region
So you can pass a gridded data file for the nest as the "mask_file"
rather
than having to create a set of lat/lon points. I'd suggest going that
route.
Here's an example of that using data from the MET tarball:
cd met-5.0
bin/gen_poly_mask \
data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
data/sample_fcst/2009123112/arw-fer-gep1/d01_2009123112_02400.grib \
mask.nc -name MY_MASK
Unfortunately, I don't think I fully understand your last question as
to why
you're getting slightly different matched pair counts (621, 619 and
617). But
my guess is one of two possibilities. Either...
(1) The polyline mask for the different grids is producing slightly
different
results and the differences lie along the boundary of the mask.
(2) There's some missing data values somewhere in the forecast and
observations causing slightly different matched pairs.
I'd suggest configuring point_stat to dump out the MPR lines for those
three
runs. Then we could take a close look at them to see where the
differences
lie. Identifying the stations where the differences occur is the
first step
in finding an explanation.
Thanks,
John
On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
> I tried another method to define the masking region using the
> gen_poly_mask tool using my input forecast file and a poly file
which
> defines the rectangular subdomain in the center of the two larger
> input forecast domains and was successful, but the results I get are
> no different than when I use the poly text file itself. This brings
up
> another question about the results I get with the poly text file. If
I
> run Point-Stat with three different input forecast grids (all
> generated from the same WRF run over a set of nested domains which
> have a common center) with three different resolutions (grid
spacing)
> and specify the same masking region using the text poly file which
> defines the domain of the innermost nest, the three CNT output files
> show three different values (621, 619 and 617) for the TOTAL number
of
> matched pairs for the variable TMP/Z2. Why is this the case when the
> scoring is being done over the same masking region? My expectation
was
> that the TOTAL number of matched pairs should all be the same.
>
> Thanks.
>
> R/
> John
>
> -----Original Message-----
> From: Raby, John W CIV USARMY ARL (US)
> Sent: Thursday, March 12, 2015 1:17 PM
> To: met_help at ucar.edu
> Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> UPDATED info
>
> I was able to locate some info in the README file in the software
> distribution folder which describes the entries in the config file
to
> describe the field in the forecast file to be used. I modified my
> config file as follows:
>
> mask = {
> grid = [];
> poly =
>
>
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> 2-07_1
> 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> ];
> sid = [];
> };
>
> Point-Stat executed for a longer period, but no output. Had an ERROR
> as
> follows:
>
> ERROR :
> ERROR : parse_poly_mask() -> the masking grid does not match the
> input grid.
> ERROR :
>
> This appears to present an issue with the way I want to use the
mask_grid.
> My intent was to score the input forecast file, but only over a
> smaller domain centered inside the domain of the input forecast
file.
> Any thoughts on this development and how I can proceed?
>
> Thanks.
> John
> _______________________________________
> From: met_help at ucar.edu via RT [met_help at ucar.edu]
> Sent: Thursday, March 12, 2015 11:44 AM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
mask_grid
> issue
>
> Greetings,
>
> This message has been automatically generated in response to the
> creation of a trouble ticket regarding:
> "MET Point-Stat mask_grid issue", a summary of which appears
below.
>
> There is no need to reply to this message right now. Your ticket
has
> been assigned an ID of [rt.rap.ucar.edu #71078].
>
> Please include the string:
>
> [rt.rap.ucar.edu #71078]
>
> in the subject line of all future correspondence about this issue.
To
> do so, you may reply to this message.
>
> Thank you,
> met_help at ucar.edu
>
>
----------------------------------------------------------------------
> --- I'm using MET V4.1 and tried to specify use of a forecast grid
> (GRIB1
> format) to define a verification masking region per the guidance on
> the User's Guide on p. 4-12 ("...any gridded data file that MET can
read..."
> .
> I
> used the following entry in the config file:
>
> mask = {
> grid =
>
>
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> 2-07_1
> 2:00:00"];
> poly = [];
> sid = [];
> };
>
> The ERROR dialog is as follows:
>
> DEBUG 4: parse_grid_mask() -> parsing grid mask
>
> "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02
> -07_12
> :00:00"
> ERROR :
> ERROR : parse_grid_mask() -> the mask_grid requested
>
> "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02
> -07_12
> :00:00" is not defined.
> ERROR :
>
> The User's Guide mentions that the user must specify a description
of
> the field to be used from the input file, but I wasn't sure how that
> info is input.
>
> Do you have any suggestions I can try to resolve this issue?
>
> Thanks.
>
> John Raby
> US Army Research Laboratory
>
>
>
>
>
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
>
>
>
Classification: UNCLASSIFIED
Caveats: NONE
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: John Halley Gotway
Time: Fri Mar 13 11:25:50 2015
John,
Thanks for letting me, and yes, please let me know what your testing
reveals.
I apologize the for user's guide being out-of-date and/or unclear.
We'll
be rewriting that section when we replace gen_poly_mask with
gen_vx_mask
anyway. But we do have larger issues with the maintenance of our
user's
guide. We are hoping/planning to make changes in how we do our
documentation to make it much easier to keep it in line with the code
itself.
Thanks,
John
On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
> John -
>
> Thanks. Your answers address all my questions very well.
>
> My lat/lon polyline has only four points so I concur that the
runtime is a
> non-issue for sure.
>
> I am very interested in your suggestion about using the
gen_poly_mask tool
> and
> providing the GRIB file of the smaller forecast as the input
"mask_file". I
> just generated the usage in an xterm and see the wording you show
below. I
> should have checked there. Maybe an edit change for the User's Guide
p.
> 3-24
> (V4.1 or later update) would also help.
>
> Regarding the different matched pair counts, I have the MPR files
for the 3
> runs and could do a "diff" among them and some other exploring to
look for
> the
> differences. I will let you know what differences I find. I suspect
that
> the
> first possibility you mention below is the more likely explanation.
If the
> input forecast grid is referenced when converting the polygon points
into
> the
> x/y values and the same with the observation locations, then that
x/y grid
> space might differ some from that which would be referenced when a
forecast
> grid with closer grid spacing is used with the same polygon points
and
> observations.
>
> I am hoping that by using the gen_poly_mask tool and providing the
GRIB
> file
> of the smaller forecast as the input "mask_file", a more accurate
> replication
> of the desired masking region might result which results in more
> consistency
> in the numbers of matched pairs. I will let you know the outcome.
>
> R/
> John
>
> -----Original Message-----
> From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> Sent: Friday, March 13, 2015 10:30 AM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED
> info (UNCLASSIFIED)
>
> John,
>
> Glad you were able to make some progress. I see a couple of
remaining
> questions in there...
>
> "Using the output of gen_poly_mask and the lat/lon polyline itself
produces
> the same result. Is that OK?"
>
> My answer is... that's great, and I'm relieved they do! The main
reason to
> use gen_poly_mask is to make the MET statistics tools (i.e.
point_stat or
> grid_stat) run a bit faster. It can be slow to figure out which
points are
> inside/outside a polyline region if the polyline contains thousands
of
> point.
> In that case, you'd run gen_poly_mask to create the mask once rather
than
> having to recreate it each time you run a MET statistics tool.
> But if your polyline only contains a small number of points, running
> gen_poly_mask first would only save you a second or two.
>
> Next, you were trying to use a smaller grid to create a mask for a
larger
> grid. The gen_poly_mask tool does support this. By the way, since
> METv5.0,
> we've realized that gen-"poly"-mask is poorly named, and in the next
> release
> it'll be replaced by a more generalized tool named gen_vx_mask. It
include
> some more flexible options.
>
> But if you read the usage statement for gen_poly_mask, the...
> "mask_file" is an ASCII Lat/Lon polyline file or gridded data
file
> defining
> the masking region
>
> So you can pass a gridded data file for the nest as the "mask_file"
rather
> than having to create a set of lat/lon points. I'd suggest going
that
> route.
>
> Here's an example of that using data from the MET tarball:
> cd met-5.0
> bin/gen_poly_mask \
> data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> data/sample_fcst/2009123112/arw-fer-gep1/d01_2009123112_02400.grib \
> mask.nc -name MY_MASK
>
> Unfortunately, I don't think I fully understand your last question
as to
> why
> you're getting slightly different matched pair counts (621, 619 and
617).
> But
> my guess is one of two possibilities. Either...
> (1) The polyline mask for the different grids is producing slightly
> different
> results and the differences lie along the boundary of the mask.
> (2) There's some missing data values somewhere in the forecast and
> observations causing slightly different matched pairs.
>
> I'd suggest configuring point_stat to dump out the MPR lines for
those
> three
> runs. Then we could take a close look at them to see where the
differences
> lie. Identifying the stations where the differences occur is the
first
> step
> in finding an explanation.
>
> Thanks,
> John
>
>
>
> On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> > I tried another method to define the masking region using the
> > gen_poly_mask tool using my input forecast file and a poly file
which
> > defines the rectangular subdomain in the center of the two larger
> > input forecast domains and was successful, but the results I get
are
> > no different than when I use the poly text file itself. This
brings up
> > another question about the results I get with the poly text file.
If I
> > run Point-Stat with three different input forecast grids (all
> > generated from the same WRF run over a set of nested domains which
> > have a common center) with three different resolutions (grid
spacing)
> > and specify the same masking region using the text poly file which
> > defines the domain of the innermost nest, the three CNT output
files
> > show three different values (621, 619 and 617) for the TOTAL
number of
> > matched pairs for the variable TMP/Z2. Why is this the case when
the
> > scoring is being done over the same masking region? My expectation
was
> > that the TOTAL number of matched pairs should all be the same.
> >
> > Thanks.
> >
> > R/
> > John
> >
> > -----Original Message-----
> > From: Raby, John W CIV USARMY ARL (US)
> > Sent: Thursday, March 12, 2015 1:17 PM
> > To: met_help at ucar.edu
> > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> > UPDATED info
> >
> > I was able to locate some info in the README file in the software
> > distribution folder which describes the entries in the config file
to
> > describe the field in the forecast file to be used. I modified my
> > config file as follows:
> >
> > mask = {
> > grid = [];
> > poly =
> >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > 2-07_1
> > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > ];
> > sid = [];
> > };
> >
> > Point-Stat executed for a longer period, but no output. Had an
ERROR
> > as
> > follows:
> >
> > ERROR :
> > ERROR : parse_poly_mask() -> the masking grid does not match the
> > input grid.
> > ERROR :
> >
> > This appears to present an issue with the way I want to use the
> mask_grid.
> > My intent was to score the input forecast file, but only over a
> > smaller domain centered inside the domain of the input forecast
file.
> > Any thoughts on this development and how I can proceed?
> >
> > Thanks.
> > John
> > _______________________________________
> > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > Sent: Thursday, March 12, 2015 11:44 AM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
mask_grid
> > issue
> >
> > Greetings,
> >
> > This message has been automatically generated in response to the
> > creation of a trouble ticket regarding:
> > "MET Point-Stat mask_grid issue", a summary of which
appears
> below.
> >
> > There is no need to reply to this message right now. Your ticket
has
> > been assigned an ID of [rt.rap.ucar.edu #71078].
> >
> > Please include the string:
> >
> > [rt.rap.ucar.edu #71078]
> >
> > in the subject line of all future correspondence about this issue.
To
> > do so, you may reply to this message.
> >
> > Thank you,
> > met_help at ucar.edu
> >
> >
----------------------------------------------------------------------
> > --- I'm using MET V4.1 and tried to specify use of a forecast grid
> > (GRIB1
> > format) to define a verification masking region per the guidance
on
> > the User's Guide on p. 4-12 ("...any gridded data file that MET
can
> read..."
> > .
> > I
> > used the following entry in the config file:
> >
> > mask = {
> > grid =
> >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > 2-07_1
> > 2:00:00"];
> > poly = [];
> > sid = [];
> > };
> >
> > The ERROR dialog is as follows:
> >
> > DEBUG 4: parse_grid_mask() -> parsing grid mask
> >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > -07_12
> > :00:00"
> > ERROR :
> > ERROR : parse_grid_mask() -> the mask_grid requested
> >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > -07_12
> > :00:00" is not defined.
> > ERROR :
> >
> > The User's Guide mentions that the user must specify a description
of
> > the field to be used from the input file, but I wasn't sure how
that
> > info is input.
> >
> > Do you have any suggestions I can try to resolve this issue?
> >
> > Thanks.
> >
> > John Raby
> > US Army Research Laboratory
> >
> >
> >
> >
> >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> >
> >
> >
>
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
>
>
>
------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more UPDATED info (UNCLASSIFIED)
From: Raby, John W USA CIV
Time: Fri Mar 13 12:17:58 2015
John -
Just ran the following command:
gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
LAX_Dumais_D1_d03_poly.nc
Received an error as follows:
ERROR :
ERROR : write_netcdf() -> the masking variable name must be set using
the "-name" command line argument.
ERROR :
The output .nc file was generated. What is the masking variable? The
usage says that the -name is optional. Is the output OK?
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Friday, March 13, 2015 11:25 AM
To: Raby, John W CIV USARMY ARL (US)
Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more
UPDATED info (UNCLASSIFIED)
John,
Thanks for letting me, and yes, please let me know what your testing
reveals.
I apologize the for user's guide being out-of-date and/or unclear.
We'll
be rewriting that section when we replace gen_poly_mask with
gen_vx_mask
anyway. But we do have larger issues with the maintenance of our
user's
guide. We are hoping/planning to make changes in how we do our
documentation to make it much easier to keep it in line with the code
itself.
Thanks,
John
On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
> John -
>
> Thanks. Your answers address all my questions very well.
>
> My lat/lon polyline has only four points so I concur that the
runtime is a
> non-issue for sure.
>
> I am very interested in your suggestion about using the
gen_poly_mask tool
> and
> providing the GRIB file of the smaller forecast as the input
"mask_file". I
> just generated the usage in an xterm and see the wording you show
below. I
> should have checked there. Maybe an edit change for the User's Guide
p.
> 3-24
> (V4.1 or later update) would also help.
>
> Regarding the different matched pair counts, I have the MPR files
for the 3
> runs and could do a "diff" among them and some other exploring to
look for
> the
> differences. I will let you know what differences I find. I suspect
that
> the
> first possibility you mention below is the more likely explanation.
If the
> input forecast grid is referenced when converting the polygon points
into
> the
> x/y values and the same with the observation locations, then that
x/y grid
> space might differ some from that which would be referenced when a
forecast
> grid with closer grid spacing is used with the same polygon points
and
> observations.
>
> I am hoping that by using the gen_poly_mask tool and providing the
GRIB
> file
> of the smaller forecast as the input "mask_file", a more accurate
> replication
> of the desired masking region might result which results in more
> consistency
> in the numbers of matched pairs. I will let you know the outcome.
>
> R/
> John
>
> -----Original Message-----
> From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> Sent: Friday, March 13, 2015 10:30 AM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED
> info (UNCLASSIFIED)
>
> John,
>
> Glad you were able to make some progress. I see a couple of
remaining
> questions in there...
>
> "Using the output of gen_poly_mask and the lat/lon polyline itself
produces
> the same result. Is that OK?"
>
> My answer is... that's great, and I'm relieved they do! The main
reason to
> use gen_poly_mask is to make the MET statistics tools (i.e.
point_stat or
> grid_stat) run a bit faster. It can be slow to figure out which
points are
> inside/outside a polyline region if the polyline contains thousands
of
> point.
> In that case, you'd run gen_poly_mask to create the mask once rather
than
> having to recreate it each time you run a MET statistics tool.
> But if your polyline only contains a small number of points, running
> gen_poly_mask first would only save you a second or two.
>
> Next, you were trying to use a smaller grid to create a mask for a
larger
> grid. The gen_poly_mask tool does support this. By the way, since
> METv5.0,
> we've realized that gen-"poly"-mask is poorly named, and in the next
> release
> it'll be replaced by a more generalized tool named gen_vx_mask. It
include
> some more flexible options.
>
> But if you read the usage statement for gen_poly_mask, the...
> "mask_file" is an ASCII Lat/Lon polyline file or gridded data
file
> defining
> the masking region
>
> So you can pass a gridded data file for the nest as the "mask_file"
rather
> than having to create a set of lat/lon points. I'd suggest going
that
> route.
>
> Here's an example of that using data from the MET tarball:
> cd met-5.0
> bin/gen_poly_mask \
> data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> data/sample_fcst/2009123112/arw-fer-gep1/d01_2009123112_02400.grib \
> mask.nc -name MY_MASK
>
> Unfortunately, I don't think I fully understand your last question
as to
> why
> you're getting slightly different matched pair counts (621, 619 and
617).
> But
> my guess is one of two possibilities. Either...
> (1) The polyline mask for the different grids is producing slightly
> different
> results and the differences lie along the boundary of the mask.
> (2) There's some missing data values somewhere in the forecast and
> observations causing slightly different matched pairs.
>
> I'd suggest configuring point_stat to dump out the MPR lines for
those
> three
> runs. Then we could take a close look at them to see where the
differences
> lie. Identifying the stations where the differences occur is the
first
> step
> in finding an explanation.
>
> Thanks,
> John
>
>
>
> On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> > I tried another method to define the masking region using the
> > gen_poly_mask tool using my input forecast file and a poly file
which
> > defines the rectangular subdomain in the center of the two larger
> > input forecast domains and was successful, but the results I get
are
> > no different than when I use the poly text file itself. This
brings up
> > another question about the results I get with the poly text file.
If I
> > run Point-Stat with three different input forecast grids (all
> > generated from the same WRF run over a set of nested domains which
> > have a common center) with three different resolutions (grid
spacing)
> > and specify the same masking region using the text poly file which
> > defines the domain of the innermost nest, the three CNT output
files
> > show three different values (621, 619 and 617) for the TOTAL
number of
> > matched pairs for the variable TMP/Z2. Why is this the case when
the
> > scoring is being done over the same masking region? My expectation
was
> > that the TOTAL number of matched pairs should all be the same.
> >
> > Thanks.
> >
> > R/
> > John
> >
> > -----Original Message-----
> > From: Raby, John W CIV USARMY ARL (US)
> > Sent: Thursday, March 12, 2015 1:17 PM
> > To: met_help at ucar.edu
> > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> > UPDATED info
> >
> > I was able to locate some info in the README file in the software
> > distribution folder which describes the entries in the config file
to
> > describe the field in the forecast file to be used. I modified my
> > config file as follows:
> >
> > mask = {
> > grid = [];
> > poly =
> >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > 2-07_1
> > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > ];
> > sid = [];
> > };
> >
> > Point-Stat executed for a longer period, but no output. Had an
ERROR
> > as
> > follows:
> >
> > ERROR :
> > ERROR : parse_poly_mask() -> the masking grid does not match the
> > input grid.
> > ERROR :
> >
> > This appears to present an issue with the way I want to use the
> mask_grid.
> > My intent was to score the input forecast file, but only over a
> > smaller domain centered inside the domain of the input forecast
file.
> > Any thoughts on this development and how I can proceed?
> >
> > Thanks.
> > John
> > _______________________________________
> > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > Sent: Thursday, March 12, 2015 11:44 AM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
mask_grid
> > issue
> >
> > Greetings,
> >
> > This message has been automatically generated in response to the
> > creation of a trouble ticket regarding:
> > "MET Point-Stat mask_grid issue", a summary of which
appears
> below.
> >
> > There is no need to reply to this message right now. Your ticket
has
> > been assigned an ID of [rt.rap.ucar.edu #71078].
> >
> > Please include the string:
> >
> > [rt.rap.ucar.edu #71078]
> >
> > in the subject line of all future correspondence about this issue.
To
> > do so, you may reply to this message.
> >
> > Thank you,
> > met_help at ucar.edu
> >
> >
----------------------------------------------------------------------
> > --- I'm using MET V4.1 and tried to specify use of a forecast grid
> > (GRIB1
> > format) to define a verification masking region per the guidance
on
> > the User's Guide on p. 4-12 ("...any gridded data file that MET
can
> read..."
> > .
> > I
> > used the following entry in the config file:
> >
> > mask = {
> > grid =
> >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > 2-07_1
> > 2:00:00"];
> > poly = [];
> > sid = [];
> > };
> >
> > The ERROR dialog is as follows:
> >
> > DEBUG 4: parse_grid_mask() -> parsing grid mask
> >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > -07_12
> > :00:00"
> > ERROR :
> > ERROR : parse_grid_mask() -> the mask_grid requested
> >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > -07_12
> > :00:00" is not defined.
> > ERROR :
> >
> > The User's Guide mentions that the user must specify a description
of
> > the field to be used from the input file, but I wasn't sure how
that
> > info is input.
> >
> > Do you have any suggestions I can try to resolve this issue?
> >
> > Thanks.
> >
> > John Raby
> > US Army Research Laboratory
> >
> >
> >
> >
> >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> >
> >
> >
>
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
>
>
>
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: John Halley Gotway
Time: Fri Mar 13 12:26:04 2015
John,
When you pass a lat/lon polyline file to gen_poly_mask, the first line
of
that polyline file contains a "name" for the masking region. That
name is
used as the variable name in the NetCDF output file. However, when
you use
a gridded data file to define the masking region, we have no good
default
value for what name you'd like to assign to that mask. That error
message
is instructing you to use the "-name" command line switch to define
it.
You might consider "-name NEST" or perhaps "-name D03_Grid".
Something
like that.
When you use the gen_poly_mask output in Point-Stat, it'll extract
that
name from the mask file and write it to the VX_MASK column in the
Point-Stat output files. So just choose whatever name you'd like to
appear
in the Point-Stat output files.
Make sense?
Thanks,
John
On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> John -
>
> Just ran the following command:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00
> \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> LAX_Dumais_D1_d03_poly.nc
>
> Received an error as follows:
>
> ERROR :
> ERROR : write_netcdf() -> the masking variable name must be set
using the
> "-name" command line argument.
> ERROR :
>
> The output .nc file was generated. What is the masking variable? The
usage
> says that the -name is optional. Is the output OK?
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Friday, March 13, 2015 11:25 AM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED info (UNCLASSIFIED)
>
> John,
>
> Thanks for letting me, and yes, please let me know what your testing
> reveals.
>
> I apologize the for user's guide being out-of-date and/or unclear.
We'll
> be rewriting that section when we replace gen_poly_mask with
gen_vx_mask
> anyway. But we do have larger issues with the maintenance of our
user's
> guide. We are hoping/planning to make changes in how we do our
> documentation to make it much easier to keep it in line with the
code
> itself.
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> > John -
> >
> > Thanks. Your answers address all my questions very well.
> >
> > My lat/lon polyline has only four points so I concur that the
runtime is
> a
> > non-issue for sure.
> >
> > I am very interested in your suggestion about using the
gen_poly_mask
> tool
> > and
> > providing the GRIB file of the smaller forecast as the input
> "mask_file". I
> > just generated the usage in an xterm and see the wording you show
below.
> I
> > should have checked there. Maybe an edit change for the User's
Guide p.
> > 3-24
> > (V4.1 or later update) would also help.
> >
> > Regarding the different matched pair counts, I have the MPR files
for
> the 3
> > runs and could do a "diff" among them and some other exploring to
look
> for
> > the
> > differences. I will let you know what differences I find. I
suspect that
> > the
> > first possibility you mention below is the more likely
explanation. If
> the
> > input forecast grid is referenced when converting the polygon
points into
> > the
> > x/y values and the same with the observation locations, then that
x/y
> grid
> > space might differ some from that which would be referenced when a
> forecast
> > grid with closer grid spacing is used with the same polygon points
and
> > observations.
> >
> > I am hoping that by using the gen_poly_mask tool and providing the
GRIB
> > file
> > of the smaller forecast as the input "mask_file", a more accurate
> > replication
> > of the desired masking region might result which results in more
> > consistency
> > in the numbers of matched pairs. I will let you know the outcome.
> >
> > R/
> > John
> >
> > -----Original Message-----
> > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 10:30 AM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > UPDATED
> > info (UNCLASSIFIED)
> >
> > John,
> >
> > Glad you were able to make some progress. I see a couple of
remaining
> > questions in there...
> >
> > "Using the output of gen_poly_mask and the lat/lon polyline itself
> produces
> > the same result. Is that OK?"
> >
> > My answer is... that's great, and I'm relieved they do! The main
reason
> to
> > use gen_poly_mask is to make the MET statistics tools (i.e.
point_stat or
> > grid_stat) run a bit faster. It can be slow to figure out which
points
> are
> > inside/outside a polyline region if the polyline contains
thousands of
> > point.
> > In that case, you'd run gen_poly_mask to create the mask once
rather than
> > having to recreate it each time you run a MET statistics tool.
> > But if your polyline only contains a small number of points,
running
> > gen_poly_mask first would only save you a second or two.
> >
> > Next, you were trying to use a smaller grid to create a mask for a
larger
> > grid. The gen_poly_mask tool does support this. By the way,
since
> > METv5.0,
> > we've realized that gen-"poly"-mask is poorly named, and in the
next
> > release
> > it'll be replaced by a more generalized tool named gen_vx_mask.
It
> include
> > some more flexible options.
> >
> > But if you read the usage statement for gen_poly_mask, the...
> > "mask_file" is an ASCII Lat/Lon polyline file or gridded data
file
> > defining
> > the masking region
> >
> > So you can pass a gridded data file for the nest as the
"mask_file"
> rather
> > than having to create a set of lat/lon points. I'd suggest going
that
> > route.
> >
> > Here's an example of that using data from the MET tarball:
> > cd met-5.0
> > bin/gen_poly_mask \
> > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > data/sample_fcst/2009123112/arw-fer-gep1/d01_2009123112_02400.grib
\
> > mask.nc -name MY_MASK
> >
> > Unfortunately, I don't think I fully understand your last question
as to
> > why
> > you're getting slightly different matched pair counts (621, 619
and 617).
> > But
> > my guess is one of two possibilities. Either...
> > (1) The polyline mask for the different grids is producing
slightly
> > different
> > results and the differences lie along the boundary of the mask.
> > (2) There's some missing data values somewhere in the forecast and
> > observations causing slightly different matched pairs.
> >
> > I'd suggest configuring point_stat to dump out the MPR lines for
those
> > three
> > runs. Then we could take a close look at them to see where the
> differences
> > lie. Identifying the stations where the differences occur is the
first
> > step
> > in finding an explanation.
> >
> > Thanks,
> > John
> >
> >
> >
> > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > > I tried another method to define the masking region using the
> > > gen_poly_mask tool using my input forecast file and a poly file
which
> > > defines the rectangular subdomain in the center of the two
larger
> > > input forecast domains and was successful, but the results I get
are
> > > no different than when I use the poly text file itself. This
brings up
> > > another question about the results I get with the poly text
file. If I
> > > run Point-Stat with three different input forecast grids (all
> > > generated from the same WRF run over a set of nested domains
which
> > > have a common center) with three different resolutions (grid
spacing)
> > > and specify the same masking region using the text poly file
which
> > > defines the domain of the innermost nest, the three CNT output
files
> > > show three different values (621, 619 and 617) for the TOTAL
number of
> > > matched pairs for the variable TMP/Z2. Why is this the case when
the
> > > scoring is being done over the same masking region? My
expectation was
> > > that the TOTAL number of matched pairs should all be the same.
> > >
> > > Thanks.
> > >
> > > R/
> > > John
> > >
> > > -----Original Message-----
> > > From: Raby, John W CIV USARMY ARL (US)
> > > Sent: Thursday, March 12, 2015 1:17 PM
> > > To: met_help at ucar.edu
> > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> > > UPDATED info
> > >
> > > I was able to locate some info in the README file in the
software
> > > distribution folder which describes the entries in the config
file to
> > > describe the field in the forecast file to be used. I modified
my
> > > config file as follows:
> > >
> > > mask = {
> > > grid = [];
> > > poly =
> > >
> > >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > 2-07_1
> > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > ];
> > > sid = [];
> > > };
> > >
> > > Point-Stat executed for a longer period, but no output. Had an
ERROR
> > > as
> > > follows:
> > >
> > > ERROR :
> > > ERROR : parse_poly_mask() -> the masking grid does not match
the
> > > input grid.
> > > ERROR :
> > >
> > > This appears to present an issue with the way I want to use the
> > mask_grid.
> > > My intent was to score the input forecast file, but only over a
> > > smaller domain centered inside the domain of the input forecast
file.
> > > Any thoughts on this development and how I can proceed?
> > >
> > > Thanks.
> > > John
> > > _______________________________________
> > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > Sent: Thursday, March 12, 2015 11:44 AM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
mask_grid
> > > issue
> > >
> > > Greetings,
> > >
> > > This message has been automatically generated in response to the
> > > creation of a trouble ticket regarding:
> > > "MET Point-Stat mask_grid issue", a summary of which
appears
> > below.
> > >
> > > There is no need to reply to this message right now. Your
ticket has
> > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > >
> > > Please include the string:
> > >
> > > [rt.rap.ucar.edu #71078]
> > >
> > > in the subject line of all future correspondence about this
issue. To
> > > do so, you may reply to this message.
> > >
> > > Thank you,
> > > met_help at ucar.edu
> > >
> > >
----------------------------------------------------------------------
> > > --- I'm using MET V4.1 and tried to specify use of a forecast
grid
> > > (GRIB1
> > > format) to define a verification masking region per the guidance
on
> > > the User's Guide on p. 4-12 ("...any gridded data file that MET
can
> > read..."
> > > .
> > > I
> > > used the following entry in the config file:
> > >
> > > mask = {
> > > grid =
> > >
> > >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > 2-07_1
> > > 2:00:00"];
> > > poly = [];
> > > sid = [];
> > > };
> > >
> > > The ERROR dialog is as follows:
> > >
> > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > >
> > >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > -07_12
> > > :00:00"
> > > ERROR :
> > > ERROR : parse_grid_mask() -> the mask_grid requested
> > >
> > >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > -07_12
> > > :00:00" is not defined.
> > > ERROR :
> > >
> > > The User's Guide mentions that the user must specify a
description of
> > > the field to be used from the input file, but I wasn't sure how
that
> > > info is input.
> > >
> > > Do you have any suggestions I can try to resolve this issue?
> > >
> > > Thanks.
> > >
> > > John Raby
> > > US Army Research Laboratory
> > >
> > >
> > >
> > >
> > >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > >
> > >
> > >
> >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more UPDATED info (UNCLASSIFIED)
From: Raby, John W USA CIV
Time: Fri Mar 13 12:30:13 2015
John -
Thanks for you help on that. I'll try your suggestion. I just tried to
visualize the output .nc file with IDV and it can't open it, which is
not surprising.
Will let you know.
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Friday, March 13, 2015 12:26 PM
To: Raby, John W CIV USARMY ARL (US)
Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more
UPDATED info (UNCLASSIFIED)
John,
When you pass a lat/lon polyline file to gen_poly_mask, the first line
of
that polyline file contains a "name" for the masking region. That
name is
used as the variable name in the NetCDF output file. However, when
you use
a gridded data file to define the masking region, we have no good
default
value for what name you'd like to assign to that mask. That error
message
is instructing you to use the "-name" command line switch to define
it.
You might consider "-name NEST" or perhaps "-name D03_Grid".
Something
like that.
When you use the gen_poly_mask output in Point-Stat, it'll extract
that
name from the mask file and write it to the VX_MASK column in the
Point-Stat output files. So just choose whatever name you'd like to
appear
in the Point-Stat output files.
Make sense?
Thanks,
John
On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> John -
>
> Just ran the following command:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00
> \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> LAX_Dumais_D1_d03_poly.nc
>
> Received an error as follows:
>
> ERROR :
> ERROR : write_netcdf() -> the masking variable name must be set
using the
> "-name" command line argument.
> ERROR :
>
> The output .nc file was generated. What is the masking variable? The
usage
> says that the -name is optional. Is the output OK?
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Friday, March 13, 2015 11:25 AM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED info (UNCLASSIFIED)
>
> John,
>
> Thanks for letting me, and yes, please let me know what your testing
> reveals.
>
> I apologize the for user's guide being out-of-date and/or unclear.
We'll
> be rewriting that section when we replace gen_poly_mask with
gen_vx_mask
> anyway. But we do have larger issues with the maintenance of our
user's
> guide. We are hoping/planning to make changes in how we do our
> documentation to make it much easier to keep it in line with the
code
> itself.
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> > John -
> >
> > Thanks. Your answers address all my questions very well.
> >
> > My lat/lon polyline has only four points so I concur that the
runtime is
> a
> > non-issue for sure.
> >
> > I am very interested in your suggestion about using the
gen_poly_mask
> tool
> > and
> > providing the GRIB file of the smaller forecast as the input
> "mask_file". I
> > just generated the usage in an xterm and see the wording you show
below.
> I
> > should have checked there. Maybe an edit change for the User's
Guide p.
> > 3-24
> > (V4.1 or later update) would also help.
> >
> > Regarding the different matched pair counts, I have the MPR files
for
> the 3
> > runs and could do a "diff" among them and some other exploring to
look
> for
> > the
> > differences. I will let you know what differences I find. I
suspect that
> > the
> > first possibility you mention below is the more likely
explanation. If
> the
> > input forecast grid is referenced when converting the polygon
points into
> > the
> > x/y values and the same with the observation locations, then that
x/y
> grid
> > space might differ some from that which would be referenced when a
> forecast
> > grid with closer grid spacing is used with the same polygon points
and
> > observations.
> >
> > I am hoping that by using the gen_poly_mask tool and providing the
GRIB
> > file
> > of the smaller forecast as the input "mask_file", a more accurate
> > replication
> > of the desired masking region might result which results in more
> > consistency
> > in the numbers of matched pairs. I will let you know the outcome.
> >
> > R/
> > John
> >
> > -----Original Message-----
> > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 10:30 AM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > UPDATED
> > info (UNCLASSIFIED)
> >
> > John,
> >
> > Glad you were able to make some progress. I see a couple of
remaining
> > questions in there...
> >
> > "Using the output of gen_poly_mask and the lat/lon polyline itself
> produces
> > the same result. Is that OK?"
> >
> > My answer is... that's great, and I'm relieved they do! The main
reason
> to
> > use gen_poly_mask is to make the MET statistics tools (i.e.
point_stat or
> > grid_stat) run a bit faster. It can be slow to figure out which
points
> are
> > inside/outside a polyline region if the polyline contains
thousands of
> > point.
> > In that case, you'd run gen_poly_mask to create the mask once
rather than
> > having to recreate it each time you run a MET statistics tool.
> > But if your polyline only contains a small number of points,
running
> > gen_poly_mask first would only save you a second or two.
> >
> > Next, you were trying to use a smaller grid to create a mask for a
larger
> > grid. The gen_poly_mask tool does support this. By the way,
since
> > METv5.0,
> > we've realized that gen-"poly"-mask is poorly named, and in the
next
> > release
> > it'll be replaced by a more generalized tool named gen_vx_mask.
It
> include
> > some more flexible options.
> >
> > But if you read the usage statement for gen_poly_mask, the...
> > "mask_file" is an ASCII Lat/Lon polyline file or gridded data
file
> > defining
> > the masking region
> >
> > So you can pass a gridded data file for the nest as the
"mask_file"
> rather
> > than having to create a set of lat/lon points. I'd suggest going
that
> > route.
> >
> > Here's an example of that using data from the MET tarball:
> > cd met-5.0
> > bin/gen_poly_mask \
> > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > data/sample_fcst/2009123112/arw-fer-gep1/d01_2009123112_02400.grib
\
> > mask.nc -name MY_MASK
> >
> > Unfortunately, I don't think I fully understand your last question
as to
> > why
> > you're getting slightly different matched pair counts (621, 619
and 617).
> > But
> > my guess is one of two possibilities. Either...
> > (1) The polyline mask for the different grids is producing
slightly
> > different
> > results and the differences lie along the boundary of the mask.
> > (2) There's some missing data values somewhere in the forecast and
> > observations causing slightly different matched pairs.
> >
> > I'd suggest configuring point_stat to dump out the MPR lines for
those
> > three
> > runs. Then we could take a close look at them to see where the
> differences
> > lie. Identifying the stations where the differences occur is the
first
> > step
> > in finding an explanation.
> >
> > Thanks,
> > John
> >
> >
> >
> > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > > I tried another method to define the masking region using the
> > > gen_poly_mask tool using my input forecast file and a poly file
which
> > > defines the rectangular subdomain in the center of the two
larger
> > > input forecast domains and was successful, but the results I get
are
> > > no different than when I use the poly text file itself. This
brings up
> > > another question about the results I get with the poly text
file. If I
> > > run Point-Stat with three different input forecast grids (all
> > > generated from the same WRF run over a set of nested domains
which
> > > have a common center) with three different resolutions (grid
spacing)
> > > and specify the same masking region using the text poly file
which
> > > defines the domain of the innermost nest, the three CNT output
files
> > > show three different values (621, 619 and 617) for the TOTAL
number of
> > > matched pairs for the variable TMP/Z2. Why is this the case when
the
> > > scoring is being done over the same masking region? My
expectation was
> > > that the TOTAL number of matched pairs should all be the same.
> > >
> > > Thanks.
> > >
> > > R/
> > > John
> > >
> > > -----Original Message-----
> > > From: Raby, John W CIV USARMY ARL (US)
> > > Sent: Thursday, March 12, 2015 1:17 PM
> > > To: met_help at ucar.edu
> > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> > > UPDATED info
> > >
> > > I was able to locate some info in the README file in the
software
> > > distribution folder which describes the entries in the config
file to
> > > describe the field in the forecast file to be used. I modified
my
> > > config file as follows:
> > >
> > > mask = {
> > > grid = [];
> > > poly =
> > >
> > >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > 2-07_1
> > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > ];
> > > sid = [];
> > > };
> > >
> > > Point-Stat executed for a longer period, but no output. Had an
ERROR
> > > as
> > > follows:
> > >
> > > ERROR :
> > > ERROR : parse_poly_mask() -> the masking grid does not match
the
> > > input grid.
> > > ERROR :
> > >
> > > This appears to present an issue with the way I want to use the
> > mask_grid.
> > > My intent was to score the input forecast file, but only over a
> > > smaller domain centered inside the domain of the input forecast
file.
> > > Any thoughts on this development and how I can proceed?
> > >
> > > Thanks.
> > > John
> > > _______________________________________
> > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > Sent: Thursday, March 12, 2015 11:44 AM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
mask_grid
> > > issue
> > >
> > > Greetings,
> > >
> > > This message has been automatically generated in response to the
> > > creation of a trouble ticket regarding:
> > > "MET Point-Stat mask_grid issue", a summary of which
appears
> > below.
> > >
> > > There is no need to reply to this message right now. Your
ticket has
> > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > >
> > > Please include the string:
> > >
> > > [rt.rap.ucar.edu #71078]
> > >
> > > in the subject line of all future correspondence about this
issue. To
> > > do so, you may reply to this message.
> > >
> > > Thank you,
> > > met_help at ucar.edu
> > >
> > >
----------------------------------------------------------------------
> > > --- I'm using MET V4.1 and tried to specify use of a forecast
grid
> > > (GRIB1
> > > format) to define a verification masking region per the guidance
on
> > > the User's Guide on p. 4-12 ("...any gridded data file that MET
can
> > read..."
> > > .
> > > I
> > > used the following entry in the config file:
> > >
> > > mask = {
> > > grid =
> > >
> > >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > 2-07_1
> > > 2:00:00"];
> > > poly = [];
> > > sid = [];
> > > };
> > >
> > > The ERROR dialog is as follows:
> > >
> > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > >
> > >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > -07_12
> > > :00:00"
> > > ERROR :
> > > ERROR : parse_grid_mask() -> the mask_grid requested
> > >
> > >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > -07_12
> > > :00:00" is not defined.
> > > ERROR :
> > >
> > > The User's Guide mentions that the user must specify a
description of
> > > the field to be used from the input file, but I wasn't sure how
that
> > > info is input.
> > >
> > > Do you have any suggestions I can try to resolve this issue?
> > >
> > > Thanks.
> > >
> > > John Raby
> > > US Army Research Laboratory
> > >
> > >
> > >
> > >
> > >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > >
> > >
> > >
> >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more UPDATED info (UNCLASSIFIED)
From: Raby, John W USA CIV
Time: Fri Mar 13 12:55:54 2015
John -
I ran the following command:
gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
131127A_WRFPRS_d03_2012-02-07_12:00:00 \
-name LAX_Dumais_D3
but, no output is generated and all that gets printed to the screen is
the usage.
I then ran:
gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
131127A_WRFPRS_d03_2012-02-07_12:00:00 \
-name LAX_Dumais_D3 -v 3
To see if the verbosity would identify the problem, but again no
output was generated and only the usage was printed to the screen.
I then tried:
gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
131127A_WRFPRS_d03_2012-02-07_12:00:00 \
-name LAX_Dumais_D3.nc V3
Any explanatio for what is happening here?
This time output was generated with name "V3" and when I looked at V3
in IDV, the title of the graphic which shows my masking region is
"LAX_Dumais_D3". The graphic look correct. Can I use this same mask
file in combination with other input forecast files as long as that
area is contained within the domain of the forecast files?
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Friday, March 13, 2015 12:26 PM
To: Raby, John W CIV USARMY ARL (US)
Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more
UPDATED info (UNCLASSIFIED)
John,
When you pass a lat/lon polyline file to gen_poly_mask, the first line
of
that polyline file contains a "name" for the masking region. That
name is
used as the variable name in the NetCDF output file. However, when
you use
a gridded data file to define the masking region, we have no good
default
value for what name you'd like to assign to that mask. That error
message
is instructing you to use the "-name" command line switch to define
it.
You might consider "-name NEST" or perhaps "-name D03_Grid".
Something
like that.
When you use the gen_poly_mask output in Point-Stat, it'll extract
that
name from the mask file and write it to the VX_MASK column in the
Point-Stat output files. So just choose whatever name you'd like to
appear
in the Point-Stat output files.
Make sense?
Thanks,
John
On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> John -
>
> Just ran the following command:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00
> \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> LAX_Dumais_D1_d03_poly.nc
>
> Received an error as follows:
>
> ERROR :
> ERROR : write_netcdf() -> the masking variable name must be set
using the
> "-name" command line argument.
> ERROR :
>
> The output .nc file was generated. What is the masking variable? The
usage
> says that the -name is optional. Is the output OK?
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Friday, March 13, 2015 11:25 AM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED info (UNCLASSIFIED)
>
> John,
>
> Thanks for letting me, and yes, please let me know what your testing
> reveals.
>
> I apologize the for user's guide being out-of-date and/or unclear.
We'll
> be rewriting that section when we replace gen_poly_mask with
gen_vx_mask
> anyway. But we do have larger issues with the maintenance of our
user's
> guide. We are hoping/planning to make changes in how we do our
> documentation to make it much easier to keep it in line with the
code
> itself.
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> > John -
> >
> > Thanks. Your answers address all my questions very well.
> >
> > My lat/lon polyline has only four points so I concur that the
runtime is
> a
> > non-issue for sure.
> >
> > I am very interested in your suggestion about using the
gen_poly_mask
> tool
> > and
> > providing the GRIB file of the smaller forecast as the input
> "mask_file". I
> > just generated the usage in an xterm and see the wording you show
below.
> I
> > should have checked there. Maybe an edit change for the User's
Guide p.
> > 3-24
> > (V4.1 or later update) would also help.
> >
> > Regarding the different matched pair counts, I have the MPR files
for
> the 3
> > runs and could do a "diff" among them and some other exploring to
look
> for
> > the
> > differences. I will let you know what differences I find. I
suspect that
> > the
> > first possibility you mention below is the more likely
explanation. If
> the
> > input forecast grid is referenced when converting the polygon
points into
> > the
> > x/y values and the same with the observation locations, then that
x/y
> grid
> > space might differ some from that which would be referenced when a
> forecast
> > grid with closer grid spacing is used with the same polygon points
and
> > observations.
> >
> > I am hoping that by using the gen_poly_mask tool and providing the
GRIB
> > file
> > of the smaller forecast as the input "mask_file", a more accurate
> > replication
> > of the desired masking region might result which results in more
> > consistency
> > in the numbers of matched pairs. I will let you know the outcome.
> >
> > R/
> > John
> >
> > -----Original Message-----
> > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 10:30 AM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > UPDATED
> > info (UNCLASSIFIED)
> >
> > John,
> >
> > Glad you were able to make some progress. I see a couple of
remaining
> > questions in there...
> >
> > "Using the output of gen_poly_mask and the lat/lon polyline itself
> produces
> > the same result. Is that OK?"
> >
> > My answer is... that's great, and I'm relieved they do! The main
reason
> to
> > use gen_poly_mask is to make the MET statistics tools (i.e.
point_stat or
> > grid_stat) run a bit faster. It can be slow to figure out which
points
> are
> > inside/outside a polyline region if the polyline contains
thousands of
> > point.
> > In that case, you'd run gen_poly_mask to create the mask once
rather than
> > having to recreate it each time you run a MET statistics tool.
> > But if your polyline only contains a small number of points,
running
> > gen_poly_mask first would only save you a second or two.
> >
> > Next, you were trying to use a smaller grid to create a mask for a
larger
> > grid. The gen_poly_mask tool does support this. By the way,
since
> > METv5.0,
> > we've realized that gen-"poly"-mask is poorly named, and in the
next
> > release
> > it'll be replaced by a more generalized tool named gen_vx_mask.
It
> include
> > some more flexible options.
> >
> > But if you read the usage statement for gen_poly_mask, the...
> > "mask_file" is an ASCII Lat/Lon polyline file or gridded data
file
> > defining
> > the masking region
> >
> > So you can pass a gridded data file for the nest as the
"mask_file"
> rather
> > than having to create a set of lat/lon points. I'd suggest going
that
> > route.
> >
> > Here's an example of that using data from the MET tarball:
> > cd met-5.0
> > bin/gen_poly_mask \
> > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > data/sample_fcst/2009123112/arw-fer-gep1/d01_2009123112_02400.grib
\
> > mask.nc -name MY_MASK
> >
> > Unfortunately, I don't think I fully understand your last question
as to
> > why
> > you're getting slightly different matched pair counts (621, 619
and 617).
> > But
> > my guess is one of two possibilities. Either...
> > (1) The polyline mask for the different grids is producing
slightly
> > different
> > results and the differences lie along the boundary of the mask.
> > (2) There's some missing data values somewhere in the forecast and
> > observations causing slightly different matched pairs.
> >
> > I'd suggest configuring point_stat to dump out the MPR lines for
those
> > three
> > runs. Then we could take a close look at them to see where the
> differences
> > lie. Identifying the stations where the differences occur is the
first
> > step
> > in finding an explanation.
> >
> > Thanks,
> > John
> >
> >
> >
> > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > > I tried another method to define the masking region using the
> > > gen_poly_mask tool using my input forecast file and a poly file
which
> > > defines the rectangular subdomain in the center of the two
larger
> > > input forecast domains and was successful, but the results I get
are
> > > no different than when I use the poly text file itself. This
brings up
> > > another question about the results I get with the poly text
file. If I
> > > run Point-Stat with three different input forecast grids (all
> > > generated from the same WRF run over a set of nested domains
which
> > > have a common center) with three different resolutions (grid
spacing)
> > > and specify the same masking region using the text poly file
which
> > > defines the domain of the innermost nest, the three CNT output
files
> > > show three different values (621, 619 and 617) for the TOTAL
number of
> > > matched pairs for the variable TMP/Z2. Why is this the case when
the
> > > scoring is being done over the same masking region? My
expectation was
> > > that the TOTAL number of matched pairs should all be the same.
> > >
> > > Thanks.
> > >
> > > R/
> > > John
> > >
> > > -----Original Message-----
> > > From: Raby, John W CIV USARMY ARL (US)
> > > Sent: Thursday, March 12, 2015 1:17 PM
> > > To: met_help at ucar.edu
> > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> > > UPDATED info
> > >
> > > I was able to locate some info in the README file in the
software
> > > distribution folder which describes the entries in the config
file to
> > > describe the field in the forecast file to be used. I modified
my
> > > config file as follows:
> > >
> > > mask = {
> > > grid = [];
> > > poly =
> > >
> > >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > 2-07_1
> > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > ];
> > > sid = [];
> > > };
> > >
> > > Point-Stat executed for a longer period, but no output. Had an
ERROR
> > > as
> > > follows:
> > >
> > > ERROR :
> > > ERROR : parse_poly_mask() -> the masking grid does not match
the
> > > input grid.
> > > ERROR :
> > >
> > > This appears to present an issue with the way I want to use the
> > mask_grid.
> > > My intent was to score the input forecast file, but only over a
> > > smaller domain centered inside the domain of the input forecast
file.
> > > Any thoughts on this development and how I can proceed?
> > >
> > > Thanks.
> > > John
> > > _______________________________________
> > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > Sent: Thursday, March 12, 2015 11:44 AM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
mask_grid
> > > issue
> > >
> > > Greetings,
> > >
> > > This message has been automatically generated in response to the
> > > creation of a trouble ticket regarding:
> > > "MET Point-Stat mask_grid issue", a summary of which
appears
> > below.
> > >
> > > There is no need to reply to this message right now. Your
ticket has
> > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > >
> > > Please include the string:
> > >
> > > [rt.rap.ucar.edu #71078]
> > >
> > > in the subject line of all future correspondence about this
issue. To
> > > do so, you may reply to this message.
> > >
> > > Thank you,
> > > met_help at ucar.edu
> > >
> > >
----------------------------------------------------------------------
> > > --- I'm using MET V4.1 and tried to specify use of a forecast
grid
> > > (GRIB1
> > > format) to define a verification masking region per the guidance
on
> > > the User's Guide on p. 4-12 ("...any gridded data file that MET
can
> > read..."
> > > .
> > > I
> > > used the following entry in the config file:
> > >
> > > mask = {
> > > grid =
> > >
> > >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > 2-07_1
> > > 2:00:00"];
> > > poly = [];
> > > sid = [];
> > > };
> > >
> > > The ERROR dialog is as follows:
> > >
> > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > >
> > >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > -07_12
> > > :00:00"
> > > ERROR :
> > > ERROR : parse_grid_mask() -> the mask_grid requested
> > >
> > >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > -07_12
> > > :00:00" is not defined.
> > > ERROR :
> > >
> > > The User's Guide mentions that the user must specify a
description of
> > > the field to be used from the input file, but I wasn't sure how
that
> > > info is input.
> > >
> > > Do you have any suggestions I can try to resolve this issue?
> > >
> > > Thanks.
> > >
> > > John Raby
> > > US Army Research Laboratory
> > >
> > >
> > >
> > >
> > >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > >
> > >
> > >
> >
> >
> > Classification: UNCLASSIFIED
> > Caveats: NONE
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: John Halley Gotway
Time: Fri Mar 13 13:00:46 2015
John,
Just take a closer look at the usage statement for gen_poly_mask...
you
need to pass in 3 file names: input data_file name, input mask_file
name,
and the output netcdf_file name.
Try this...
gen_poly_mask \
Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
131127A_WRFPRS_d03_2012-02-07_12:00:00 \
LAX_Dumais_D3.nc \
-name LAX_MASK
Or use whatever value for the "-name" option you'd like.
Thanks,
John
On Fri, Mar 13, 2015 at 12:55 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> John -
>
> I ran the following command:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> -name LAX_Dumais_D3
>
> but, no output is generated and all that gets printed to the screen
is the
> usage.
>
> I then ran:
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> -name LAX_Dumais_D3 -v 3
>
> To see if the verbosity would identify the problem, but again no
output
> was generated and only the usage was printed to the screen.
>
> I then tried:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> -name LAX_Dumais_D3.nc V3
>
> Any explanatio for what is happening here?
>
> This time output was generated with name "V3" and when I looked at
V3 in
> IDV, the title of the graphic which shows my masking region is
> "LAX_Dumais_D3". The graphic look correct. Can I use this same mask
file in
> combination with other input forecast files as long as that area is
> contained within the domain of the forecast files?
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Friday, March 13, 2015 12:26 PM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED info (UNCLASSIFIED)
>
> John,
>
> When you pass a lat/lon polyline file to gen_poly_mask, the first
line of
> that polyline file contains a "name" for the masking region. That
name is
> used as the variable name in the NetCDF output file. However, when
you use
> a gridded data file to define the masking region, we have no good
default
> value for what name you'd like to assign to that mask. That error
message
> is instructing you to use the "-name" command line switch to define
it.
> You might consider "-name NEST" or perhaps "-name D03_Grid".
Something
> like that.
>
> When you use the gen_poly_mask output in Point-Stat, it'll extract
that
> name from the mask file and write it to the VX_MASK column in the
> Point-Stat output files. So just choose whatever name you'd like to
appear
> in the Point-Stat output files.
>
> Make sense?
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > John -
> >
> > Just ran the following command:
> >
> > gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00
> > \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > LAX_Dumais_D1_d03_poly.nc
> >
> > Received an error as follows:
> >
> > ERROR :
> > ERROR : write_netcdf() -> the masking variable name must be set
using
> the
> > "-name" command line argument.
> > ERROR :
> >
> > The output .nc file was generated. What is the masking variable?
The
> usage
> > says that the -name is optional. Is the output OK?
> >
> > R/
> > John
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 11:25 AM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > UPDATED info (UNCLASSIFIED)
> >
> > John,
> >
> > Thanks for letting me, and yes, please let me know what your
testing
> > reveals.
> >
> > I apologize the for user's guide being out-of-date and/or unclear.
We'll
> > be rewriting that section when we replace gen_poly_mask with
gen_vx_mask
> > anyway. But we do have larger issues with the maintenance of our
user's
> > guide. We are hoping/planning to make changes in how we do our
> > documentation to make it much easier to keep it in line with the
code
> > itself.
> >
> > Thanks,
> > John
> >
> > On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > > John -
> > >
> > > Thanks. Your answers address all my questions very well.
> > >
> > > My lat/lon polyline has only four points so I concur that the
runtime
> is
> > a
> > > non-issue for sure.
> > >
> > > I am very interested in your suggestion about using the
gen_poly_mask
> > tool
> > > and
> > > providing the GRIB file of the smaller forecast as the input
> > "mask_file". I
> > > just generated the usage in an xterm and see the wording you
show
> below.
> > I
> > > should have checked there. Maybe an edit change for the User's
Guide p.
> > > 3-24
> > > (V4.1 or later update) would also help.
> > >
> > > Regarding the different matched pair counts, I have the MPR
files for
> > the 3
> > > runs and could do a "diff" among them and some other exploring
to look
> > for
> > > the
> > > differences. I will let you know what differences I find. I
suspect
> that
> > > the
> > > first possibility you mention below is the more likely
explanation. If
> > the
> > > input forecast grid is referenced when converting the polygon
points
> into
> > > the
> > > x/y values and the same with the observation locations, then
that x/y
> > grid
> > > space might differ some from that which would be referenced when
a
> > forecast
> > > grid with closer grid spacing is used with the same polygon
points and
> > > observations.
> > >
> > > I am hoping that by using the gen_poly_mask tool and providing
the GRIB
> > > file
> > > of the smaller forecast as the input "mask_file", a more
accurate
> > > replication
> > > of the desired masking region might result which results in more
> > > consistency
> > > in the numbers of matched pairs. I will let you know the
outcome.
> > >
> > > R/
> > > John
> > >
> > > -----Original Message-----
> > > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > > Sent: Friday, March 13, 2015 10:30 AM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > > UPDATED
> > > info (UNCLASSIFIED)
> > >
> > > John,
> > >
> > > Glad you were able to make some progress. I see a couple of
remaining
> > > questions in there...
> > >
> > > "Using the output of gen_poly_mask and the lat/lon polyline
itself
> > produces
> > > the same result. Is that OK?"
> > >
> > > My answer is... that's great, and I'm relieved they do! The
main
> reason
> > to
> > > use gen_poly_mask is to make the MET statistics tools (i.e.
point_stat
> or
> > > grid_stat) run a bit faster. It can be slow to figure out which
points
> > are
> > > inside/outside a polyline region if the polyline contains
thousands of
> > > point.
> > > In that case, you'd run gen_poly_mask to create the mask once
rather
> than
> > > having to recreate it each time you run a MET statistics tool.
> > > But if your polyline only contains a small number of points,
running
> > > gen_poly_mask first would only save you a second or two.
> > >
> > > Next, you were trying to use a smaller grid to create a mask for
a
> larger
> > > grid. The gen_poly_mask tool does support this. By the way,
since
> > > METv5.0,
> > > we've realized that gen-"poly"-mask is poorly named, and in the
next
> > > release
> > > it'll be replaced by a more generalized tool named gen_vx_mask.
It
> > include
> > > some more flexible options.
> > >
> > > But if you read the usage statement for gen_poly_mask, the...
> > > "mask_file" is an ASCII Lat/Lon polyline file or gridded data
file
> > > defining
> > > the masking region
> > >
> > > So you can pass a gridded data file for the nest as the
"mask_file"
> > rather
> > > than having to create a set of lat/lon points. I'd suggest
going that
> > > route.
> > >
> > > Here's an example of that using data from the MET tarball:
> > > cd met-5.0
> > > bin/gen_poly_mask \
> > > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > > data/sample_fcst/2009123112/arw-fer-
gep1/d01_2009123112_02400.grib \
> > > mask.nc -name MY_MASK
> > >
> > > Unfortunately, I don't think I fully understand your last
question as
> to
> > > why
> > > you're getting slightly different matched pair counts (621, 619
and
> 617).
> > > But
> > > my guess is one of two possibilities. Either...
> > > (1) The polyline mask for the different grids is producing
slightly
> > > different
> > > results and the differences lie along the boundary of the mask.
> > > (2) There's some missing data values somewhere in the forecast
and
> > > observations causing slightly different matched pairs.
> > >
> > > I'd suggest configuring point_stat to dump out the MPR lines for
those
> > > three
> > > runs. Then we could take a close look at them to see where the
> > differences
> > > lie. Identifying the stations where the differences occur is
the first
> > > step
> > > in finding an explanation.
> > >
> > > Thanks,
> > > John
> > >
> > >
> > >
> > > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078
>
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > > I tried another method to define the masking region using the
> > > > gen_poly_mask tool using my input forecast file and a poly
file which
> > > > defines the rectangular subdomain in the center of the two
larger
> > > > input forecast domains and was successful, but the results I
get are
> > > > no different than when I use the poly text file itself. This
brings
> up
> > > > another question about the results I get with the poly text
file. If
> I
> > > > run Point-Stat with three different input forecast grids (all
> > > > generated from the same WRF run over a set of nested domains
which
> > > > have a common center) with three different resolutions (grid
spacing)
> > > > and specify the same masking region using the text poly file
which
> > > > defines the domain of the innermost nest, the three CNT output
files
> > > > show three different values (621, 619 and 617) for the TOTAL
number
> of
> > > > matched pairs for the variable TMP/Z2. Why is this the case
when the
> > > > scoring is being done over the same masking region? My
expectation
> was
> > > > that the TOTAL number of matched pairs should all be the same.
> > > >
> > > > Thanks.
> > > >
> > > > R/
> > > > John
> > > >
> > > > -----Original Message-----
> > > > From: Raby, John W CIV USARMY ARL (US)
> > > > Sent: Thursday, March 12, 2015 1:17 PM
> > > > To: met_help at ucar.edu
> > > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> > > > UPDATED info
> > > >
> > > > I was able to locate some info in the README file in the
software
> > > > distribution folder which describes the entries in the config
file to
> > > > describe the field in the forecast file to be used. I modified
my
> > > > config file as follows:
> > > >
> > > > mask = {
> > > > grid = [];
> > > > poly =
> > > >
> > > >
>
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > > 2-07_1
> > > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > > ];
> > > > sid = [];
> > > > };
> > > >
> > > > Point-Stat executed for a longer period, but no output. Had an
ERROR
> > > > as
> > > > follows:
> > > >
> > > > ERROR :
> > > > ERROR : parse_poly_mask() -> the masking grid does not match
the
> > > > input grid.
> > > > ERROR :
> > > >
> > > > This appears to present an issue with the way I want to use
the
> > > mask_grid.
> > > > My intent was to score the input forecast file, but only over
a
> > > > smaller domain centered inside the domain of the input
forecast file.
> > > > Any thoughts on this development and how I can proceed?
> > > >
> > > > Thanks.
> > > > John
> > > > _______________________________________
> > > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > > Sent: Thursday, March 12, 2015 11:44 AM
> > > > To: Raby, John W CIV USARMY ARL (US)
> > > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
> mask_grid
> > > > issue
> > > >
> > > > Greetings,
> > > >
> > > > This message has been automatically generated in response to
the
> > > > creation of a trouble ticket regarding:
> > > > "MET Point-Stat mask_grid issue", a summary of which
appears
> > > below.
> > > >
> > > > There is no need to reply to this message right now. Your
ticket has
> > > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > > >
> > > > Please include the string:
> > > >
> > > > [rt.rap.ucar.edu #71078]
> > > >
> > > > in the subject line of all future correspondence about this
issue. To
> > > > do so, you may reply to this message.
> > > >
> > > > Thank you,
> > > > met_help at ucar.edu
> > > >
> > > >
>
----------------------------------------------------------------------
> > > > --- I'm using MET V4.1 and tried to specify use of a forecast
grid
> > > > (GRIB1
> > > > format) to define a verification masking region per the
guidance on
> > > > the User's Guide on p. 4-12 ("...any gridded data file that
MET can
> > > read..."
> > > > .
> > > > I
> > > > used the following entry in the config file:
> > > >
> > > > mask = {
> > > > grid =
> > > >
> > > >
>
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > > 2-07_1
> > > > 2:00:00"];
> > > > poly = [];
> > > > sid = [];
> > > > };
> > > >
> > > > The ERROR dialog is as follows:
> > > >
> > > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > > >
> > > >
> "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02
> > > > -07_12
> > > > :00:00"
> > > > ERROR :
> > > > ERROR : parse_grid_mask() -> the mask_grid requested
> > > >
> > > >
> "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02
> > > > -07_12
> > > > :00:00" is not defined.
> > > > ERROR :
> > > >
> > > > The User's Guide mentions that the user must specify a
description of
> > > > the field to be used from the input file, but I wasn't sure
how that
> > > > info is input.
> > > >
> > > > Do you have any suggestions I can try to resolve this issue?
> > > >
> > > > Thanks.
> > > >
> > > > John Raby
> > > > US Army Research Laboratory
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > >
> > >
> > >
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more UPDATED info (UNCLASSIFIED)
From: Raby, John W USA CIV
Time: Fri Mar 13 13:17:34 2015
John -
Thanks. that correction fixed it.
Just ran:
gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
131127A_WRFPRS_d03_2012-02-07_12:00:00 \
LAX_Dumais_D1_d03_poly.nc \
-name DumaisD3mask
With IDV, I visualize the file: LAX_Dumais_D2_d03_poly.nc and the
title of the graphic is DumaisD3mask
Can this same .nc file be used as input to Point-Stat where I've used
other input forecast grids as long as the forecast grid contains the
area defined in the .nc file?
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Friday, March 13, 2015 1:00 PM
To: Raby, John W CIV USARMY ARL (US)
Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more
UPDATED info (UNCLASSIFIED)
John,
Just take a closer look at the usage statement for gen_poly_mask...
you
need to pass in 3 file names: input data_file name, input mask_file
name,
and the output netcdf_file name.
Try this...
gen_poly_mask \
Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
131127A_WRFPRS_d03_2012-02-07_12:00:00 \
LAX_Dumais_D3.nc \
-name LAX_MASK
Or use whatever value for the "-name" option you'd like.
Thanks,
John
On Fri, Mar 13, 2015 at 12:55 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> John -
>
> I ran the following command:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> -name LAX_Dumais_D3
>
> but, no output is generated and all that gets printed to the screen
is the
> usage.
>
> I then ran:
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> -name LAX_Dumais_D3 -v 3
>
> To see if the verbosity would identify the problem, but again no
output
> was generated and only the usage was printed to the screen.
>
> I then tried:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> -name LAX_Dumais_D3.nc V3
>
> Any explanatio for what is happening here?
>
> This time output was generated with name "V3" and when I looked at
V3 in
> IDV, the title of the graphic which shows my masking region is
> "LAX_Dumais_D3". The graphic look correct. Can I use this same mask
file in
> combination with other input forecast files as long as that area is
> contained within the domain of the forecast files?
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Friday, March 13, 2015 12:26 PM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED info (UNCLASSIFIED)
>
> John,
>
> When you pass a lat/lon polyline file to gen_poly_mask, the first
line of
> that polyline file contains a "name" for the masking region. That
name is
> used as the variable name in the NetCDF output file. However, when
you use
> a gridded data file to define the masking region, we have no good
default
> value for what name you'd like to assign to that mask. That error
message
> is instructing you to use the "-name" command line switch to define
it.
> You might consider "-name NEST" or perhaps "-name D03_Grid".
Something
> like that.
>
> When you use the gen_poly_mask output in Point-Stat, it'll extract
that
> name from the mask file and write it to the VX_MASK column in the
> Point-Stat output files. So just choose whatever name you'd like to
appear
> in the Point-Stat output files.
>
> Make sense?
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > John -
> >
> > Just ran the following command:
> >
> > gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00
> > \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > LAX_Dumais_D1_d03_poly.nc
> >
> > Received an error as follows:
> >
> > ERROR :
> > ERROR : write_netcdf() -> the masking variable name must be set
using
> the
> > "-name" command line argument.
> > ERROR :
> >
> > The output .nc file was generated. What is the masking variable?
The
> usage
> > says that the -name is optional. Is the output OK?
> >
> > R/
> > John
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 11:25 AM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > UPDATED info (UNCLASSIFIED)
> >
> > John,
> >
> > Thanks for letting me, and yes, please let me know what your
testing
> > reveals.
> >
> > I apologize the for user's guide being out-of-date and/or unclear.
We'll
> > be rewriting that section when we replace gen_poly_mask with
gen_vx_mask
> > anyway. But we do have larger issues with the maintenance of our
user's
> > guide. We are hoping/planning to make changes in how we do our
> > documentation to make it much easier to keep it in line with the
code
> > itself.
> >
> > Thanks,
> > John
> >
> > On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > > John -
> > >
> > > Thanks. Your answers address all my questions very well.
> > >
> > > My lat/lon polyline has only four points so I concur that the
runtime
> is
> > a
> > > non-issue for sure.
> > >
> > > I am very interested in your suggestion about using the
gen_poly_mask
> > tool
> > > and
> > > providing the GRIB file of the smaller forecast as the input
> > "mask_file". I
> > > just generated the usage in an xterm and see the wording you
show
> below.
> > I
> > > should have checked there. Maybe an edit change for the User's
Guide p.
> > > 3-24
> > > (V4.1 or later update) would also help.
> > >
> > > Regarding the different matched pair counts, I have the MPR
files for
> > the 3
> > > runs and could do a "diff" among them and some other exploring
to look
> > for
> > > the
> > > differences. I will let you know what differences I find. I
suspect
> that
> > > the
> > > first possibility you mention below is the more likely
explanation. If
> > the
> > > input forecast grid is referenced when converting the polygon
points
> into
> > > the
> > > x/y values and the same with the observation locations, then
that x/y
> > grid
> > > space might differ some from that which would be referenced when
a
> > forecast
> > > grid with closer grid spacing is used with the same polygon
points and
> > > observations.
> > >
> > > I am hoping that by using the gen_poly_mask tool and providing
the GRIB
> > > file
> > > of the smaller forecast as the input "mask_file", a more
accurate
> > > replication
> > > of the desired masking region might result which results in more
> > > consistency
> > > in the numbers of matched pairs. I will let you know the
outcome.
> > >
> > > R/
> > > John
> > >
> > > -----Original Message-----
> > > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > > Sent: Friday, March 13, 2015 10:30 AM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > > UPDATED
> > > info (UNCLASSIFIED)
> > >
> > > John,
> > >
> > > Glad you were able to make some progress. I see a couple of
remaining
> > > questions in there...
> > >
> > > "Using the output of gen_poly_mask and the lat/lon polyline
itself
> > produces
> > > the same result. Is that OK?"
> > >
> > > My answer is... that's great, and I'm relieved they do! The
main
> reason
> > to
> > > use gen_poly_mask is to make the MET statistics tools (i.e.
point_stat
> or
> > > grid_stat) run a bit faster. It can be slow to figure out which
points
> > are
> > > inside/outside a polyline region if the polyline contains
thousands of
> > > point.
> > > In that case, you'd run gen_poly_mask to create the mask once
rather
> than
> > > having to recreate it each time you run a MET statistics tool.
> > > But if your polyline only contains a small number of points,
running
> > > gen_poly_mask first would only save you a second or two.
> > >
> > > Next, you were trying to use a smaller grid to create a mask for
a
> larger
> > > grid. The gen_poly_mask tool does support this. By the way,
since
> > > METv5.0,
> > > we've realized that gen-"poly"-mask is poorly named, and in the
next
> > > release
> > > it'll be replaced by a more generalized tool named gen_vx_mask.
It
> > include
> > > some more flexible options.
> > >
> > > But if you read the usage statement for gen_poly_mask, the...
> > > "mask_file" is an ASCII Lat/Lon polyline file or gridded data
file
> > > defining
> > > the masking region
> > >
> > > So you can pass a gridded data file for the nest as the
"mask_file"
> > rather
> > > than having to create a set of lat/lon points. I'd suggest
going that
> > > route.
> > >
> > > Here's an example of that using data from the MET tarball:
> > > cd met-5.0
> > > bin/gen_poly_mask \
> > > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > > data/sample_fcst/2009123112/arw-fer-
gep1/d01_2009123112_02400.grib \
> > > mask.nc -name MY_MASK
> > >
> > > Unfortunately, I don't think I fully understand your last
question as
> to
> > > why
> > > you're getting slightly different matched pair counts (621, 619
and
> 617).
> > > But
> > > my guess is one of two possibilities. Either...
> > > (1) The polyline mask for the different grids is producing
slightly
> > > different
> > > results and the differences lie along the boundary of the mask.
> > > (2) There's some missing data values somewhere in the forecast
and
> > > observations causing slightly different matched pairs.
> > >
> > > I'd suggest configuring point_stat to dump out the MPR lines for
those
> > > three
> > > runs. Then we could take a close look at them to see where the
> > differences
> > > lie. Identifying the stations where the differences occur is
the first
> > > step
> > > in finding an explanation.
> > >
> > > Thanks,
> > > John
> > >
> > >
> > >
> > > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078
>
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > > I tried another method to define the masking region using the
> > > > gen_poly_mask tool using my input forecast file and a poly
file which
> > > > defines the rectangular subdomain in the center of the two
larger
> > > > input forecast domains and was successful, but the results I
get are
> > > > no different than when I use the poly text file itself. This
brings
> up
> > > > another question about the results I get with the poly text
file. If
> I
> > > > run Point-Stat with three different input forecast grids (all
> > > > generated from the same WRF run over a set of nested domains
which
> > > > have a common center) with three different resolutions (grid
spacing)
> > > > and specify the same masking region using the text poly file
which
> > > > defines the domain of the innermost nest, the three CNT output
files
> > > > show three different values (621, 619 and 617) for the TOTAL
number
> of
> > > > matched pairs for the variable TMP/Z2. Why is this the case
when the
> > > > scoring is being done over the same masking region? My
expectation
> was
> > > > that the TOTAL number of matched pairs should all be the same.
> > > >
> > > > Thanks.
> > > >
> > > > R/
> > > > John
> > > >
> > > > -----Original Message-----
> > > > From: Raby, John W CIV USARMY ARL (US)
> > > > Sent: Thursday, March 12, 2015 1:17 PM
> > > > To: met_help at ucar.edu
> > > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> > > > UPDATED info
> > > >
> > > > I was able to locate some info in the README file in the
software
> > > > distribution folder which describes the entries in the config
file to
> > > > describe the field in the forecast file to be used. I modified
my
> > > > config file as follows:
> > > >
> > > > mask = {
> > > > grid = [];
> > > > poly =
> > > >
> > > >
>
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > > 2-07_1
> > > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > > ];
> > > > sid = [];
> > > > };
> > > >
> > > > Point-Stat executed for a longer period, but no output. Had an
ERROR
> > > > as
> > > > follows:
> > > >
> > > > ERROR :
> > > > ERROR : parse_poly_mask() -> the masking grid does not match
the
> > > > input grid.
> > > > ERROR :
> > > >
> > > > This appears to present an issue with the way I want to use
the
> > > mask_grid.
> > > > My intent was to score the input forecast file, but only over
a
> > > > smaller domain centered inside the domain of the input
forecast file.
> > > > Any thoughts on this development and how I can proceed?
> > > >
> > > > Thanks.
> > > > John
> > > > _______________________________________
> > > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > > Sent: Thursday, March 12, 2015 11:44 AM
> > > > To: Raby, John W CIV USARMY ARL (US)
> > > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
> mask_grid
> > > > issue
> > > >
> > > > Greetings,
> > > >
> > > > This message has been automatically generated in response to
the
> > > > creation of a trouble ticket regarding:
> > > > "MET Point-Stat mask_grid issue", a summary of which
appears
> > > below.
> > > >
> > > > There is no need to reply to this message right now. Your
ticket has
> > > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > > >
> > > > Please include the string:
> > > >
> > > > [rt.rap.ucar.edu #71078]
> > > >
> > > > in the subject line of all future correspondence about this
issue. To
> > > > do so, you may reply to this message.
> > > >
> > > > Thank you,
> > > > met_help at ucar.edu
> > > >
> > > >
>
----------------------------------------------------------------------
> > > > --- I'm using MET V4.1 and tried to specify use of a forecast
grid
> > > > (GRIB1
> > > > format) to define a verification masking region per the
guidance on
> > > > the User's Guide on p. 4-12 ("...any gridded data file that
MET can
> > > read..."
> > > > .
> > > > I
> > > > used the following entry in the config file:
> > > >
> > > > mask = {
> > > > grid =
> > > >
> > > >
>
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > > 2-07_1
> > > > 2:00:00"];
> > > > poly = [];
> > > > sid = [];
> > > > };
> > > >
> > > > The ERROR dialog is as follows:
> > > >
> > > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > > >
> > > >
> "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02
> > > > -07_12
> > > > :00:00"
> > > > ERROR :
> > > > ERROR : parse_grid_mask() -> the mask_grid requested
> > > >
> > > >
> "/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
02
> > > > -07_12
> > > > :00:00" is not defined.
> > > > ERROR :
> > > >
> > > > The User's Guide mentions that the user must specify a
description of
> > > > the field to be used from the input file, but I wasn't sure
how that
> > > > info is input.
> > > >
> > > > Do you have any suggestions I can try to resolve this issue?
> > > >
> > > > Thanks.
> > > >
> > > > John Raby
> > > > US Army Research Laboratory
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > Classification: UNCLASSIFIED
> > > Caveats: NONE
> > >
> > >
> > >
> > >
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: John Halley Gotway
Time: Fri Mar 13 13:36:54 2015
John,
When you run Point-Stat, it reads the grid information from the input
forecast file. That grid definition includes the number of x/y
points,
lat/lon origin, and several other parameters. When it reads the
gridded
masking file, it also extracts the grid information from it. If those
2
grid definitions differ in any way, Point-Stat will error out. So you
need
to run gen_poly_mask for each forecast grid you want to evaluate.
Thanks,
John
On Fri, Mar 13, 2015 at 1:17 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> John -
> Thanks. that correction fixed it.
> Just ran:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> LAX_Dumais_D1_d03_poly.nc \
> -name DumaisD3mask
>
> With IDV, I visualize the file: LAX_Dumais_D2_d03_poly.nc and the
title of
> the graphic is DumaisD3mask
>
> Can this same .nc file be used as input to Point-Stat where I've
used
> other input forecast grids as long as the forecast grid contains the
area
> defined in the .nc file?
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Friday, March 13, 2015 1:00 PM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED info (UNCLASSIFIED)
>
> John,
>
> Just take a closer look at the usage statement for gen_poly_mask...
you
> need to pass in 3 file names: input data_file name, input mask_file
name,
> and the output netcdf_file name.
>
> Try this...
>
> gen_poly_mask \
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> LAX_Dumais_D3.nc \
> -name LAX_MASK
>
> Or use whatever value for the "-name" option you'd like.
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 12:55 PM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > John -
> >
> > I ran the following command:
> >
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > -name LAX_Dumais_D3
> >
> > but, no output is generated and all that gets printed to the
screen is
> the
> > usage.
> >
> > I then ran:
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > -name LAX_Dumais_D3 -v 3
> >
> > To see if the verbosity would identify the problem, but again no
output
> > was generated and only the usage was printed to the screen.
> >
> > I then tried:
> >
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > -name LAX_Dumais_D3.nc V3
> >
> > Any explanatio for what is happening here?
> >
> > This time output was generated with name "V3" and when I looked at
V3 in
> > IDV, the title of the graphic which shows my masking region is
> > "LAX_Dumais_D3". The graphic look correct. Can I use this same
mask file
> in
> > combination with other input forecast files as long as that area
is
> > contained within the domain of the forecast files?
> >
> > R/
> > John
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 12:26 PM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > UPDATED info (UNCLASSIFIED)
> >
> > John,
> >
> > When you pass a lat/lon polyline file to gen_poly_mask, the first
line of
> > that polyline file contains a "name" for the masking region. That
name
> is
> > used as the variable name in the NetCDF output file. However,
when you
> use
> > a gridded data file to define the masking region, we have no good
default
> > value for what name you'd like to assign to that mask. That error
> message
> > is instructing you to use the "-name" command line switch to
define it.
> > You might consider "-name NEST" or perhaps "-name D03_Grid".
Something
> > like that.
> >
> > When you use the gen_poly_mask output in Point-Stat, it'll extract
that
> > name from the mask file and write it to the VX_MASK column in the
> > Point-Stat output files. So just choose whatever name you'd like
to
> appear
> > in the Point-Stat output files.
> >
> > Make sense?
> >
> > Thanks,
> > John
> >
> > On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > John -
> > >
> > > Just ran the following command:
> > >
> > > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00
> > > \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > > LAX_Dumais_D1_d03_poly.nc
> > >
> > > Received an error as follows:
> > >
> > > ERROR :
> > > ERROR : write_netcdf() -> the masking variable name must be set
using
> > the
> > > "-name" command line argument.
> > > ERROR :
> > >
> > > The output .nc file was generated. What is the masking variable?
The
> > usage
> > > says that the -name is optional. Is the output OK?
> > >
> > > R/
> > > John
> > > ________________________________________
> > > From: John Halley Gotway via RT [met_help at ucar.edu]
> > > Sent: Friday, March 13, 2015 11:25 AM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > > UPDATED info (UNCLASSIFIED)
> > >
> > > John,
> > >
> > > Thanks for letting me, and yes, please let me know what your
testing
> > > reveals.
> > >
> > > I apologize the for user's guide being out-of-date and/or
unclear.
> We'll
> > > be rewriting that section when we replace gen_poly_mask with
> gen_vx_mask
> > > anyway. But we do have larger issues with the maintenance of
our
> user's
> > > guide. We are hoping/planning to make changes in how we do our
> > > documentation to make it much easier to keep it in line with the
code
> > > itself.
> > >
> > > Thanks,
> > > John
> > >
> > > On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078
>
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > > John -
> > > >
> > > > Thanks. Your answers address all my questions very well.
> > > >
> > > > My lat/lon polyline has only four points so I concur that the
runtime
> > is
> > > a
> > > > non-issue for sure.
> > > >
> > > > I am very interested in your suggestion about using the
gen_poly_mask
> > > tool
> > > > and
> > > > providing the GRIB file of the smaller forecast as the input
> > > "mask_file". I
> > > > just generated the usage in an xterm and see the wording you
show
> > below.
> > > I
> > > > should have checked there. Maybe an edit change for the User's
Guide
> p.
> > > > 3-24
> > > > (V4.1 or later update) would also help.
> > > >
> > > > Regarding the different matched pair counts, I have the MPR
files for
> > > the 3
> > > > runs and could do a "diff" among them and some other exploring
to
> look
> > > for
> > > > the
> > > > differences. I will let you know what differences I find. I
suspect
> > that
> > > > the
> > > > first possibility you mention below is the more likely
explanation.
> If
> > > the
> > > > input forecast grid is referenced when converting the polygon
points
> > into
> > > > the
> > > > x/y values and the same with the observation locations, then
that x/y
> > > grid
> > > > space might differ some from that which would be referenced
when a
> > > forecast
> > > > grid with closer grid spacing is used with the same polygon
points
> and
> > > > observations.
> > > >
> > > > I am hoping that by using the gen_poly_mask tool and providing
the
> GRIB
> > > > file
> > > > of the smaller forecast as the input "mask_file", a more
accurate
> > > > replication
> > > > of the desired masking region might result which results in
more
> > > > consistency
> > > > in the numbers of matched pairs. I will let you know the
outcome.
> > > >
> > > > R/
> > > > John
> > > >
> > > > -----Original Message-----
> > > > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > > > Sent: Friday, March 13, 2015 10:30 AM
> > > > To: Raby, John W CIV USARMY ARL (US)
> > > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> more
> > > > UPDATED
> > > > info (UNCLASSIFIED)
> > > >
> > > > John,
> > > >
> > > > Glad you were able to make some progress. I see a couple of
> remaining
> > > > questions in there...
> > > >
> > > > "Using the output of gen_poly_mask and the lat/lon polyline
itself
> > > produces
> > > > the same result. Is that OK?"
> > > >
> > > > My answer is... that's great, and I'm relieved they do! The
main
> > reason
> > > to
> > > > use gen_poly_mask is to make the MET statistics tools (i.e.
> point_stat
> > or
> > > > grid_stat) run a bit faster. It can be slow to figure out
which
> points
> > > are
> > > > inside/outside a polyline region if the polyline contains
thousands
> of
> > > > point.
> > > > In that case, you'd run gen_poly_mask to create the mask once
rather
> > than
> > > > having to recreate it each time you run a MET statistics tool.
> > > > But if your polyline only contains a small number of points,
running
> > > > gen_poly_mask first would only save you a second or two.
> > > >
> > > > Next, you were trying to use a smaller grid to create a mask
for a
> > larger
> > > > grid. The gen_poly_mask tool does support this. By the way,
since
> > > > METv5.0,
> > > > we've realized that gen-"poly"-mask is poorly named, and in
the next
> > > > release
> > > > it'll be replaced by a more generalized tool named
gen_vx_mask. It
> > > include
> > > > some more flexible options.
> > > >
> > > > But if you read the usage statement for gen_poly_mask, the...
> > > > "mask_file" is an ASCII Lat/Lon polyline file or gridded
data file
> > > > defining
> > > > the masking region
> > > >
> > > > So you can pass a gridded data file for the nest as the
"mask_file"
> > > rather
> > > > than having to create a set of lat/lon points. I'd suggest
going
> that
> > > > route.
> > > >
> > > > Here's an example of that using data from the MET tarball:
> > > > cd met-5.0
> > > > bin/gen_poly_mask \
> > > > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > > > data/sample_fcst/2009123112/arw-fer-
gep1/d01_2009123112_02400.grib \
> > > > mask.nc -name MY_MASK
> > > >
> > > > Unfortunately, I don't think I fully understand your last
question as
> > to
> > > > why
> > > > you're getting slightly different matched pair counts (621,
619 and
> > 617).
> > > > But
> > > > my guess is one of two possibilities. Either...
> > > > (1) The polyline mask for the different grids is producing
slightly
> > > > different
> > > > results and the differences lie along the boundary of the
mask.
> > > > (2) There's some missing data values somewhere in the forecast
and
> > > > observations causing slightly different matched pairs.
> > > >
> > > > I'd suggest configuring point_stat to dump out the MPR lines
for
> those
> > > > three
> > > > runs. Then we could take a close look at them to see where
the
> > > differences
> > > > lie. Identifying the stations where the differences occur is
the
> first
> > > > step
> > > > in finding an explanation.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > >
> > > >
> > > > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > > > >
> > > > > Classification: UNCLASSIFIED
> > > > > Caveats: NONE
> > > > >
> > > > > I tried another method to define the masking region using
the
> > > > > gen_poly_mask tool using my input forecast file and a poly
file
> which
> > > > > defines the rectangular subdomain in the center of the two
larger
> > > > > input forecast domains and was successful, but the results I
get
> are
> > > > > no different than when I use the poly text file itself. This
brings
> > up
> > > > > another question about the results I get with the poly text
file.
> If
> > I
> > > > > run Point-Stat with three different input forecast grids
(all
> > > > > generated from the same WRF run over a set of nested domains
which
> > > > > have a common center) with three different resolutions (grid
> spacing)
> > > > > and specify the same masking region using the text poly file
which
> > > > > defines the domain of the innermost nest, the three CNT
output
> files
> > > > > show three different values (621, 619 and 617) for the TOTAL
number
> > of
> > > > > matched pairs for the variable TMP/Z2. Why is this the case
when
> the
> > > > > scoring is being done over the same masking region? My
expectation
> > was
> > > > > that the TOTAL number of matched pairs should all be the
same.
> > > > >
> > > > > Thanks.
> > > > >
> > > > > R/
> > > > > John
> > > > >
> > > > > -----Original Message-----
> > > > > From: Raby, John W CIV USARMY ARL (US)
> > > > > Sent: Thursday, March 12, 2015 1:17 PM
> > > > > To: met_help at ucar.edu
> > > > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> > > > > UPDATED info
> > > > >
> > > > > I was able to locate some info in the README file in the
software
> > > > > distribution folder which describes the entries in the
config file
> to
> > > > > describe the field in the forecast file to be used. I
modified my
> > > > > config file as follows:
> > > > >
> > > > > mask = {
> > > > > grid = [];
> > > > > poly =
> > > > >
> > > > >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > > > 2-07_1
> > > > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > > > ];
> > > > > sid = [];
> > > > > };
> > > > >
> > > > > Point-Stat executed for a longer period, but no output. Had
an
> ERROR
> > > > > as
> > > > > follows:
> > > > >
> > > > > ERROR :
> > > > > ERROR : parse_poly_mask() -> the masking grid does not
match the
> > > > > input grid.
> > > > > ERROR :
> > > > >
> > > > > This appears to present an issue with the way I want to use
the
> > > > mask_grid.
> > > > > My intent was to score the input forecast file, but only
over a
> > > > > smaller domain centered inside the domain of the input
forecast
> file.
> > > > > Any thoughts on this development and how I can proceed?
> > > > >
> > > > > Thanks.
> > > > > John
> > > > > _______________________________________
> > > > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > > > Sent: Thursday, March 12, 2015 11:44 AM
> > > > > To: Raby, John W CIV USARMY ARL (US)
> > > > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
> > mask_grid
> > > > > issue
> > > > >
> > > > > Greetings,
> > > > >
> > > > > This message has been automatically generated in response to
the
> > > > > creation of a trouble ticket regarding:
> > > > > "MET Point-Stat mask_grid issue", a summary of which
> appears
> > > > below.
> > > > >
> > > > > There is no need to reply to this message right now. Your
ticket
> has
> > > > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > > > >
> > > > > Please include the string:
> > > > >
> > > > > [rt.rap.ucar.edu #71078]
> > > > >
> > > > > in the subject line of all future correspondence about this
issue.
> To
> > > > > do so, you may reply to this message.
> > > > >
> > > > > Thank you,
> > > > > met_help at ucar.edu
> > > > >
> > > > >
> >
----------------------------------------------------------------------
> > > > > --- I'm using MET V4.1 and tried to specify use of a
forecast grid
> > > > > (GRIB1
> > > > > format) to define a verification masking region per the
guidance on
> > > > > the User's Guide on p. 4-12 ("...any gridded data file that
MET can
> > > > read..."
> > > > > .
> > > > > I
> > > > > used the following entry in the config file:
> > > > >
> > > > > mask = {
> > > > > grid =
> > > > >
> > > > >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > > > 2-07_1
> > > > > 2:00:00"];
> > > > > poly = [];
> > > > > sid = [];
> > > > > };
> > > > >
> > > > > The ERROR dialog is as follows:
> > > > >
> > > > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > > > >
> > > > >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > > > -07_12
> > > > > :00:00"
> > > > > ERROR :
> > > > > ERROR : parse_grid_mask() -> the mask_grid requested
> > > > >
> > > > >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > > > -07_12
> > > > > :00:00" is not defined.
> > > > > ERROR :
> > > > >
> > > > > The User's Guide mentions that the user must specify a
description
> of
> > > > > the field to be used from the input file, but I wasn't sure
how
> that
> > > > > info is input.
> > > > >
> > > > > Do you have any suggestions I can try to resolve this issue?
> > > > >
> > > > > Thanks.
> > > > >
> > > > > John Raby
> > > > > US Army Research Laboratory
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > Classification: UNCLASSIFIED
> > > > > Caveats: NONE
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more UPDATED info (UNCLASSIFIED)
From: Raby, John W USA CIV
Time: Fri Mar 13 13:42:44 2015
John -
Thanks for answering that question. I had set up some command line
scripts which were taking that tact, but I wasn't sure, so knowing
that, I will correct them per your suggestions which resulted in
success today and proceed accordingly generating specific masking
files linked to the specific forecast grid you are evaluating. I'll
let you know the results of running Point-Stat.
R/
John
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Friday, March 13, 2015 1:36 PM
To: Raby, John W CIV USARMY ARL (US)
Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more
UPDATED info (UNCLASSIFIED)
John,
When you run Point-Stat, it reads the grid information from the input
forecast file. That grid definition includes the number of x/y
points,
lat/lon origin, and several other parameters. When it reads the
gridded
masking file, it also extracts the grid information from it. If those
2
grid definitions differ in any way, Point-Stat will error out. So you
need
to run gen_poly_mask for each forecast grid you want to evaluate.
Thanks,
John
On Fri, Mar 13, 2015 at 1:17 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> John -
> Thanks. that correction fixed it.
> Just ran:
>
> gen_poly_mask Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-
07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> LAX_Dumais_D1_d03_poly.nc \
> -name DumaisD3mask
>
> With IDV, I visualize the file: LAX_Dumais_D2_d03_poly.nc and the
title of
> the graphic is DumaisD3mask
>
> Can this same .nc file be used as input to Point-Stat where I've
used
> other input forecast grids as long as the forecast grid contains the
area
> defined in the .nc file?
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Friday, March 13, 2015 1:00 PM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED info (UNCLASSIFIED)
>
> John,
>
> Just take a closer look at the usage statement for gen_poly_mask...
you
> need to pass in 3 file names: input data_file name, input mask_file
name,
> and the output netcdf_file name.
>
> Try this...
>
> gen_poly_mask \
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> LAX_Dumais_D3.nc \
> -name LAX_MASK
>
> Or use whatever value for the "-name" option you'd like.
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 12:55 PM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > John -
> >
> > I ran the following command:
> >
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > -name LAX_Dumais_D3
> >
> > but, no output is generated and all that gets printed to the
screen is
> the
> > usage.
> >
> > I then ran:
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > -name LAX_Dumais_D3 -v 3
> >
> > To see if the verbosity would identify the problem, but again no
output
> > was generated and only the usage was printed to the screen.
> >
> > I then tried:
> >
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > -name LAX_Dumais_D3.nc V3
> >
> > Any explanatio for what is happening here?
> >
> > This time output was generated with name "V3" and when I looked at
V3 in
> > IDV, the title of the graphic which shows my masking region is
> > "LAX_Dumais_D3". The graphic look correct. Can I use this same
mask file
> in
> > combination with other input forecast files as long as that area
is
> > contained within the domain of the forecast files?
> >
> > R/
> > John
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 12:26 PM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > UPDATED info (UNCLASSIFIED)
> >
> > John,
> >
> > When you pass a lat/lon polyline file to gen_poly_mask, the first
line of
> > that polyline file contains a "name" for the masking region. That
name
> is
> > used as the variable name in the NetCDF output file. However,
when you
> use
> > a gridded data file to define the masking region, we have no good
default
> > value for what name you'd like to assign to that mask. That error
> message
> > is instructing you to use the "-name" command line switch to
define it.
> > You might consider "-name NEST" or perhaps "-name D03_Grid".
Something
> > like that.
> >
> > When you use the gen_poly_mask output in Point-Stat, it'll extract
that
> > name from the mask file and write it to the VX_MASK column in the
> > Point-Stat output files. So just choose whatever name you'd like
to
> appear
> > in the Point-Stat output files.
> >
> > Make sense?
> >
> > Thanks,
> > John
> >
> > On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > John -
> > >
> > > Just ran the following command:
> > >
> > > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00
> > > \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > > LAX_Dumais_D1_d03_poly.nc
> > >
> > > Received an error as follows:
> > >
> > > ERROR :
> > > ERROR : write_netcdf() -> the masking variable name must be set
using
> > the
> > > "-name" command line argument.
> > > ERROR :
> > >
> > > The output .nc file was generated. What is the masking variable?
The
> > usage
> > > says that the -name is optional. Is the output OK?
> > >
> > > R/
> > > John
> > > ________________________________________
> > > From: John Halley Gotway via RT [met_help at ucar.edu]
> > > Sent: Friday, March 13, 2015 11:25 AM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > > UPDATED info (UNCLASSIFIED)
> > >
> > > John,
> > >
> > > Thanks for letting me, and yes, please let me know what your
testing
> > > reveals.
> > >
> > > I apologize the for user's guide being out-of-date and/or
unclear.
> We'll
> > > be rewriting that section when we replace gen_poly_mask with
> gen_vx_mask
> > > anyway. But we do have larger issues with the maintenance of
our
> user's
> > > guide. We are hoping/planning to make changes in how we do our
> > > documentation to make it much easier to keep it in line with the
code
> > > itself.
> > >
> > > Thanks,
> > > John
> > >
> > > On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078
>
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > > John -
> > > >
> > > > Thanks. Your answers address all my questions very well.
> > > >
> > > > My lat/lon polyline has only four points so I concur that the
runtime
> > is
> > > a
> > > > non-issue for sure.
> > > >
> > > > I am very interested in your suggestion about using the
gen_poly_mask
> > > tool
> > > > and
> > > > providing the GRIB file of the smaller forecast as the input
> > > "mask_file". I
> > > > just generated the usage in an xterm and see the wording you
show
> > below.
> > > I
> > > > should have checked there. Maybe an edit change for the User's
Guide
> p.
> > > > 3-24
> > > > (V4.1 or later update) would also help.
> > > >
> > > > Regarding the different matched pair counts, I have the MPR
files for
> > > the 3
> > > > runs and could do a "diff" among them and some other exploring
to
> look
> > > for
> > > > the
> > > > differences. I will let you know what differences I find. I
suspect
> > that
> > > > the
> > > > first possibility you mention below is the more likely
explanation.
> If
> > > the
> > > > input forecast grid is referenced when converting the polygon
points
> > into
> > > > the
> > > > x/y values and the same with the observation locations, then
that x/y
> > > grid
> > > > space might differ some from that which would be referenced
when a
> > > forecast
> > > > grid with closer grid spacing is used with the same polygon
points
> and
> > > > observations.
> > > >
> > > > I am hoping that by using the gen_poly_mask tool and providing
the
> GRIB
> > > > file
> > > > of the smaller forecast as the input "mask_file", a more
accurate
> > > > replication
> > > > of the desired masking region might result which results in
more
> > > > consistency
> > > > in the numbers of matched pairs. I will let you know the
outcome.
> > > >
> > > > R/
> > > > John
> > > >
> > > > -----Original Message-----
> > > > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > > > Sent: Friday, March 13, 2015 10:30 AM
> > > > To: Raby, John W CIV USARMY ARL (US)
> > > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> more
> > > > UPDATED
> > > > info (UNCLASSIFIED)
> > > >
> > > > John,
> > > >
> > > > Glad you were able to make some progress. I see a couple of
> remaining
> > > > questions in there...
> > > >
> > > > "Using the output of gen_poly_mask and the lat/lon polyline
itself
> > > produces
> > > > the same result. Is that OK?"
> > > >
> > > > My answer is... that's great, and I'm relieved they do! The
main
> > reason
> > > to
> > > > use gen_poly_mask is to make the MET statistics tools (i.e.
> point_stat
> > or
> > > > grid_stat) run a bit faster. It can be slow to figure out
which
> points
> > > are
> > > > inside/outside a polyline region if the polyline contains
thousands
> of
> > > > point.
> > > > In that case, you'd run gen_poly_mask to create the mask once
rather
> > than
> > > > having to recreate it each time you run a MET statistics tool.
> > > > But if your polyline only contains a small number of points,
running
> > > > gen_poly_mask first would only save you a second or two.
> > > >
> > > > Next, you were trying to use a smaller grid to create a mask
for a
> > larger
> > > > grid. The gen_poly_mask tool does support this. By the way,
since
> > > > METv5.0,
> > > > we've realized that gen-"poly"-mask is poorly named, and in
the next
> > > > release
> > > > it'll be replaced by a more generalized tool named
gen_vx_mask. It
> > > include
> > > > some more flexible options.
> > > >
> > > > But if you read the usage statement for gen_poly_mask, the...
> > > > "mask_file" is an ASCII Lat/Lon polyline file or gridded
data file
> > > > defining
> > > > the masking region
> > > >
> > > > So you can pass a gridded data file for the nest as the
"mask_file"
> > > rather
> > > > than having to create a set of lat/lon points. I'd suggest
going
> that
> > > > route.
> > > >
> > > > Here's an example of that using data from the MET tarball:
> > > > cd met-5.0
> > > > bin/gen_poly_mask \
> > > > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > > > data/sample_fcst/2009123112/arw-fer-
gep1/d01_2009123112_02400.grib \
> > > > mask.nc -name MY_MASK
> > > >
> > > > Unfortunately, I don't think I fully understand your last
question as
> > to
> > > > why
> > > > you're getting slightly different matched pair counts (621,
619 and
> > 617).
> > > > But
> > > > my guess is one of two possibilities. Either...
> > > > (1) The polyline mask for the different grids is producing
slightly
> > > > different
> > > > results and the differences lie along the boundary of the
mask.
> > > > (2) There's some missing data values somewhere in the forecast
and
> > > > observations causing slightly different matched pairs.
> > > >
> > > > I'd suggest configuring point_stat to dump out the MPR lines
for
> those
> > > > three
> > > > runs. Then we could take a close look at them to see where
the
> > > differences
> > > > lie. Identifying the stations where the differences occur is
the
> first
> > > > step
> > > > in finding an explanation.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > >
> > > >
> > > > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > > > >
> > > > > Classification: UNCLASSIFIED
> > > > > Caveats: NONE
> > > > >
> > > > > I tried another method to define the masking region using
the
> > > > > gen_poly_mask tool using my input forecast file and a poly
file
> which
> > > > > defines the rectangular subdomain in the center of the two
larger
> > > > > input forecast domains and was successful, but the results I
get
> are
> > > > > no different than when I use the poly text file itself. This
brings
> > up
> > > > > another question about the results I get with the poly text
file.
> If
> > I
> > > > > run Point-Stat with three different input forecast grids
(all
> > > > > generated from the same WRF run over a set of nested domains
which
> > > > > have a common center) with three different resolutions (grid
> spacing)
> > > > > and specify the same masking region using the text poly file
which
> > > > > defines the domain of the innermost nest, the three CNT
output
> files
> > > > > show three different values (621, 619 and 617) for the TOTAL
number
> > of
> > > > > matched pairs for the variable TMP/Z2. Why is this the case
when
> the
> > > > > scoring is being done over the same masking region? My
expectation
> > was
> > > > > that the TOTAL number of matched pairs should all be the
same.
> > > > >
> > > > > Thanks.
> > > > >
> > > > > R/
> > > > > John
> > > > >
> > > > > -----Original Message-----
> > > > > From: Raby, John W CIV USARMY ARL (US)
> > > > > Sent: Thursday, March 12, 2015 1:17 PM
> > > > > To: met_help at ucar.edu
> > > > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> > > > > UPDATED info
> > > > >
> > > > > I was able to locate some info in the README file in the
software
> > > > > distribution folder which describes the entries in the
config file
> to
> > > > > describe the field in the forecast file to be used. I
modified my
> > > > > config file as follows:
> > > > >
> > > > > mask = {
> > > > > grid = [];
> > > > > poly =
> > > > >
> > > > >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > > > 2-07_1
> > > > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > > > ];
> > > > > sid = [];
> > > > > };
> > > > >
> > > > > Point-Stat executed for a longer period, but no output. Had
an
> ERROR
> > > > > as
> > > > > follows:
> > > > >
> > > > > ERROR :
> > > > > ERROR : parse_poly_mask() -> the masking grid does not
match the
> > > > > input grid.
> > > > > ERROR :
> > > > >
> > > > > This appears to present an issue with the way I want to use
the
> > > > mask_grid.
> > > > > My intent was to score the input forecast file, but only
over a
> > > > > smaller domain centered inside the domain of the input
forecast
> file.
> > > > > Any thoughts on this development and how I can proceed?
> > > > >
> > > > > Thanks.
> > > > > John
> > > > > _______________________________________
> > > > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > > > Sent: Thursday, March 12, 2015 11:44 AM
> > > > > To: Raby, John W CIV USARMY ARL (US)
> > > > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
> > mask_grid
> > > > > issue
> > > > >
> > > > > Greetings,
> > > > >
> > > > > This message has been automatically generated in response to
the
> > > > > creation of a trouble ticket regarding:
> > > > > "MET Point-Stat mask_grid issue", a summary of which
> appears
> > > > below.
> > > > >
> > > > > There is no need to reply to this message right now. Your
ticket
> has
> > > > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > > > >
> > > > > Please include the string:
> > > > >
> > > > > [rt.rap.ucar.edu #71078]
> > > > >
> > > > > in the subject line of all future correspondence about this
issue.
> To
> > > > > do so, you may reply to this message.
> > > > >
> > > > > Thank you,
> > > > > met_help at ucar.edu
> > > > >
> > > > >
> >
----------------------------------------------------------------------
> > > > > --- I'm using MET V4.1 and tried to specify use of a
forecast grid
> > > > > (GRIB1
> > > > > format) to define a verification masking region per the
guidance on
> > > > > the User's Guide on p. 4-12 ("...any gridded data file that
MET can
> > > > read..."
> > > > > .
> > > > > I
> > > > > used the following entry in the config file:
> > > > >
> > > > > mask = {
> > > > > grid =
> > > > >
> > > > >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-0
> > > > > 2-07_1
> > > > > 2:00:00"];
> > > > > poly = [];
> > > > > sid = [];
> > > > > };
> > > > >
> > > > > The ERROR dialog is as follows:
> > > > >
> > > > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > > > >
> > > > >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > > > -07_12
> > > > > :00:00"
> > > > > ERROR :
> > > > > ERROR : parse_grid_mask() -> the mask_grid requested
> > > > >
> > > > >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-02
> > > > > -07_12
> > > > > :00:00" is not defined.
> > > > > ERROR :
> > > > >
> > > > > The User's Guide mentions that the user must specify a
description
> of
> > > > > the field to be used from the input file, but I wasn't sure
how
> that
> > > > > info is input.
> > > > >
> > > > > Do you have any suggestions I can try to resolve this issue?
> > > > >
> > > > > Thanks.
> > > > >
> > > > > John Raby
> > > > > US Army Research Laboratory
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > Classification: UNCLASSIFIED
> > > > > Caveats: NONE
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
> >
> >
>
>
>
>
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: Raby, John W USA CIV
Time: Fri Mar 13 15:49:17 2015
Classification: UNCLASSIFIED
Caveats: NONE
John -
I just finished running Point-Stat with three input forecast files
each with a
different resolution (grid spacing), but for domains which are all
centered on
the same point and are all nested. I had generated the gen_poly_mask
NetCDF
masking files using the forecast file for the innermost nest to define
the
masking region. The config files for each of the three Point-Stat runs
specified the unique masking file which had been generated with that
particular forecast file and it worked fine. I addition, I checked the
total
number of matched pairs in the three Point-Stat CNT output files and
they were
identical! This is what I was trying to achieve. The VX_MASK col had
the
expected names. In IDV, the graphical depiction of the masking region
was as
expected and all three regions were coincident.
This will also me to compare the error stats for the forecasts
generated at
three different grid spacings to evaluate the value added of higher
resolution
forecasts.
Thanks for all your help.
R/
John
-----Original Message-----
From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
Sent: Friday, March 13, 2015 1:37 PM
To: Raby, John W CIV USARMY ARL (US)
Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue more
UPDATED
info (UNCLASSIFIED)
John,
When you run Point-Stat, it reads the grid information from the input
forecast
file. That grid definition includes the number of x/y points, lat/lon
origin,
and several other parameters. When it reads the gridded masking file,
it also
extracts the grid information from it. If those 2 grid definitions
differ in
any way, Point-Stat will error out. So you need to run gen_poly_mask
for each
forecast grid you want to evaluate.
Thanks,
John
On Fri, Mar 13, 2015 at 1:17 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> John -
> Thanks. that correction fixed it.
> Just ran:
>
> gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ LAX_Dumais_D1_d03_poly.nc \
> -name DumaisD3mask
>
> With IDV, I visualize the file: LAX_Dumais_D2_d03_poly.nc and the
> title of the graphic is DumaisD3mask
>
> Can this same .nc file be used as input to Point-Stat where I've
used
> other input forecast grids as long as the forecast grid contains the
> area defined in the .nc file?
>
> R/
> John
> ________________________________________
> From: John Halley Gotway via RT [met_help at ucar.edu]
> Sent: Friday, March 13, 2015 1:00 PM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED info (UNCLASSIFIED)
>
> John,
>
> Just take a closer look at the usage statement for gen_poly_mask...
> you need to pass in 3 file names: input data_file name, input
> mask_file name, and the output netcdf_file name.
>
> Try this...
>
> gen_poly_mask \
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ LAX_Dumais_D3.nc \ -name
> LAX_MASK
>
> Or use whatever value for the "-name" option you'd like.
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 12:55 PM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > John -
> >
> > I ran the following command:
> >
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ -name LAX_Dumais_D3
> >
> > but, no output is generated and all that gets printed to the
screen
> > is
> the
> > usage.
> >
> > I then ran:
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ -name LAX_Dumais_D3 -v 3
> >
> > To see if the verbosity would identify the problem, but again no
> > output was generated and only the usage was printed to the screen.
> >
> > I then tried:
> >
> > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ -name LAX_Dumais_D3.nc V3
> >
> > Any explanatio for what is happening here?
> >
> > This time output was generated with name "V3" and when I looked at
> > V3 in IDV, the title of the graphic which shows my masking region
is
> > "LAX_Dumais_D3". The graphic look correct. Can I use this same
mask
> > file
> in
> > combination with other input forecast files as long as that area
is
> > contained within the domain of the forecast files?
> >
> > R/
> > John
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 12:26 PM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> > more UPDATED info (UNCLASSIFIED)
> >
> > John,
> >
> > When you pass a lat/lon polyline file to gen_poly_mask, the first
> > line of that polyline file contains a "name" for the masking
region.
> > That name
> is
> > used as the variable name in the NetCDF output file. However,
when
> > you
> use
> > a gridded data file to define the masking region, we have no good
> > default value for what name you'd like to assign to that mask.
That
> > error
> message
> > is instructing you to use the "-name" command line switch to
define it.
> > You might consider "-name NEST" or perhaps "-name D03_Grid".
> > Something like that.
> >
> > When you use the gen_poly_mask output in Point-Stat, it'll extract
> > that name from the mask file and write it to the VX_MASK column in
> > the Point-Stat output files. So just choose whatever name you'd
> > like to
> appear
> > in the Point-Stat output files.
> >
> > Make sense?
> >
> > Thanks,
> > John
> >
> > On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > John -
> > >
> > > Just ran the following command:
> > >
> > > gen_poly_mask
> Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00
> > > \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > > LAX_Dumais_D1_d03_poly.nc
> > >
> > > Received an error as follows:
> > >
> > > ERROR :
> > > ERROR : write_netcdf() -> the masking variable name must be set
> > > using
> > the
> > > "-name" command line argument.
> > > ERROR :
> > >
> > > The output .nc file was generated. What is the masking variable?
> > > The
> > usage
> > > says that the -name is optional. Is the output OK?
> > >
> > > R/
> > > John
> > > ________________________________________
> > > From: John Halley Gotway via RT [met_help at ucar.edu]
> > > Sent: Friday, March 13, 2015 11:25 AM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> > > more UPDATED info (UNCLASSIFIED)
> > >
> > > John,
> > >
> > > Thanks for letting me, and yes, please let me know what your
> > > testing reveals.
> > >
> > > I apologize the for user's guide being out-of-date and/or
unclear.
> We'll
> > > be rewriting that section when we replace gen_poly_mask with
> gen_vx_mask
> > > anyway. But we do have larger issues with the maintenance of
our
> user's
> > > guide. We are hoping/planning to make changes in how we do our
> > > documentation to make it much easier to keep it in line with the
> > > code itself.
> > >
> > > Thanks,
> > > John
> > >
> > > On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078
>
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > > John -
> > > >
> > > > Thanks. Your answers address all my questions very well.
> > > >
> > > > My lat/lon polyline has only four points so I concur that the
> > > > runtime
> > is
> > > a
> > > > non-issue for sure.
> > > >
> > > > I am very interested in your suggestion about using the
> > > > gen_poly_mask
> > > tool
> > > > and
> > > > providing the GRIB file of the smaller forecast as the input
> > > "mask_file". I
> > > > just generated the usage in an xterm and see the wording you
> > > > show
> > below.
> > > I
> > > > should have checked there. Maybe an edit change for the User's
> > > > Guide
> p.
> > > > 3-24
> > > > (V4.1 or later update) would also help.
> > > >
> > > > Regarding the different matched pair counts, I have the MPR
> > > > files for
> > > the 3
> > > > runs and could do a "diff" among them and some other exploring
> > > > to
> look
> > > for
> > > > the
> > > > differences. I will let you know what differences I find. I
> > > > suspect
> > that
> > > > the
> > > > first possibility you mention below is the more likely
explanation.
> If
> > > the
> > > > input forecast grid is referenced when converting the polygon
> > > > points
> > into
> > > > the
> > > > x/y values and the same with the observation locations, then
> > > > that x/y
> > > grid
> > > > space might differ some from that which would be referenced
when
> > > > a
> > > forecast
> > > > grid with closer grid spacing is used with the same polygon
> > > > points
> and
> > > > observations.
> > > >
> > > > I am hoping that by using the gen_poly_mask tool and providing
> > > > the
> GRIB
> > > > file
> > > > of the smaller forecast as the input "mask_file", a more
> > > > accurate replication of the desired masking region might
result
> > > > which results in more consistency in the numbers of matched
> > > > pairs. I will let you know the outcome.
> > > >
> > > > R/
> > > > John
> > > >
> > > > -----Original Message-----
> > > > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > > > Sent: Friday, March 13, 2015 10:30 AM
> > > > To: Raby, John W CIV USARMY ARL (US)
> > > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> more
> > > > UPDATED
> > > > info (UNCLASSIFIED)
> > > >
> > > > John,
> > > >
> > > > Glad you were able to make some progress. I see a couple of
> remaining
> > > > questions in there...
> > > >
> > > > "Using the output of gen_poly_mask and the lat/lon polyline
> > > > itself
> > > produces
> > > > the same result. Is that OK?"
> > > >
> > > > My answer is... that's great, and I'm relieved they do! The
> > > > main
> > reason
> > > to
> > > > use gen_poly_mask is to make the MET statistics tools (i.e.
> point_stat
> > or
> > > > grid_stat) run a bit faster. It can be slow to figure out
which
> points
> > > are
> > > > inside/outside a polyline region if the polyline contains
> > > > thousands
> of
> > > > point.
> > > > In that case, you'd run gen_poly_mask to create the mask once
> > > > rather
> > than
> > > > having to recreate it each time you run a MET statistics tool.
> > > > But if your polyline only contains a small number of points,
> > > > running gen_poly_mask first would only save you a second or
two.
> > > >
> > > > Next, you were trying to use a smaller grid to create a mask
for
> > > > a
> > larger
> > > > grid. The gen_poly_mask tool does support this. By the way,
> > > > since METv5.0, we've realized that gen-"poly"-mask is poorly
> > > > named, and in the next release it'll be replaced by a more
> > > > generalized tool named gen_vx_mask. It
> > > include
> > > > some more flexible options.
> > > >
> > > > But if you read the usage statement for gen_poly_mask, the...
> > > > "mask_file" is an ASCII Lat/Lon polyline file or gridded
data
> > > > file defining the masking region
> > > >
> > > > So you can pass a gridded data file for the nest as the
"mask_file"
> > > rather
> > > > than having to create a set of lat/lon points. I'd suggest
> > > > going
> that
> > > > route.
> > > >
> > > > Here's an example of that using data from the MET tarball:
> > > > cd met-5.0
> > > > bin/gen_poly_mask \
> > > > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > > > data/sample_fcst/2009123112/arw-fer-
gep1/d01_2009123112_02400.gr
> > > > ib \ mask.nc -name MY_MASK
> > > >
> > > > Unfortunately, I don't think I fully understand your last
> > > > question as
> > to
> > > > why
> > > > you're getting slightly different matched pair counts (621,
619
> > > > and
> > 617).
> > > > But
> > > > my guess is one of two possibilities. Either...
> > > > (1) The polyline mask for the different grids is producing
> > > > slightly different results and the differences lie along the
> > > > boundary of the mask.
> > > > (2) There's some missing data values somewhere in the forecast
> > > > and observations causing slightly different matched pairs.
> > > >
> > > > I'd suggest configuring point_stat to dump out the MPR lines
for
> those
> > > > three
> > > > runs. Then we could take a close look at them to see where
the
> > > differences
> > > > lie. Identifying the stations where the differences occur is
> > > > the
> first
> > > > step
> > > > in finding an explanation.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > >
> > > >
> > > > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT <
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078
> > > > > >
> > > > >
> > > > > Classification: UNCLASSIFIED
> > > > > Caveats: NONE
> > > > >
> > > > > I tried another method to define the masking region using
the
> > > > > gen_poly_mask tool using my input forecast file and a poly
> > > > > file
> which
> > > > > defines the rectangular subdomain in the center of the two
> > > > > larger input forecast domains and was successful, but the
> > > > > results I get
> are
> > > > > no different than when I use the poly text file itself. This
> > > > > brings
> > up
> > > > > another question about the results I get with the poly text
file.
> If
> > I
> > > > > run Point-Stat with three different input forecast grids
(all
> > > > > generated from the same WRF run over a set of nested domains
> > > > > which have a common center) with three different resolutions
> > > > > (grid
> spacing)
> > > > > and specify the same masking region using the text poly file
> > > > > which defines the domain of the innermost nest, the three
CNT
> > > > > output
> files
> > > > > show three different values (621, 619 and 617) for the TOTAL
> > > > > number
> > of
> > > > > matched pairs for the variable TMP/Z2. Why is this the case
> > > > > when
> the
> > > > > scoring is being done over the same masking region? My
> > > > > expectation
> > was
> > > > > that the TOTAL number of matched pairs should all be the
same.
> > > > >
> > > > > Thanks.
> > > > >
> > > > > R/
> > > > > John
> > > > >
> > > > > -----Original Message-----
> > > > > From: Raby, John W CIV USARMY ARL (US)
> > > > > Sent: Thursday, March 12, 2015 1:17 PM
> > > > > To: met_help at ucar.edu
> > > > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
> > > > > issue UPDATED info
> > > > >
> > > > > I was able to locate some info in the README file in the
> > > > > software distribution folder which describes the entries in
> > > > > the config file
> to
> > > > > describe the field in the forecast file to be used. I
modified
> > > > > my config file as follows:
> > > > >
> > > > > mask = {
> > > > > grid = [];
> > > > > poly =
> > > > >
> > > > >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012
> > -0
> > > > > 2-07_1
> > > > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > > > ];
> > > > > sid = [];
> > > > > };
> > > > >
> > > > > Point-Stat executed for a longer period, but no output. Had
an
> ERROR
> > > > > as
> > > > > follows:
> > > > >
> > > > > ERROR :
> > > > > ERROR : parse_poly_mask() -> the masking grid does not
match
> > > > > the input grid.
> > > > > ERROR :
> > > > >
> > > > > This appears to present an issue with the way I want to use
> > > > > the
> > > > mask_grid.
> > > > > My intent was to score the input forecast file, but only
over
> > > > > a smaller domain centered inside the domain of the input
> > > > > forecast
> file.
> > > > > Any thoughts on this development and how I can proceed?
> > > > >
> > > > > Thanks.
> > > > > John
> > > > > _______________________________________
> > > > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > > > Sent: Thursday, March 12, 2015 11:44 AM
> > > > > To: Raby, John W CIV USARMY ARL (US)
> > > > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-Stat
> > mask_grid
> > > > > issue
> > > > >
> > > > > Greetings,
> > > > >
> > > > > This message has been automatically generated in response to
> > > > > the creation of a trouble ticket regarding:
> > > > > "MET Point-Stat mask_grid issue", a summary of which
> appears
> > > > below.
> > > > >
> > > > > There is no need to reply to this message right now. Your
> > > > > ticket
> has
> > > > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > > > >
> > > > > Please include the string:
> > > > >
> > > > > [rt.rap.ucar.edu #71078]
> > > > >
> > > > > in the subject line of all future correspondence about this
issue.
> To
> > > > > do so, you may reply to this message.
> > > > >
> > > > > Thank you,
> > > > > met_help at ucar.edu
> > > > >
> > > > >
> >
--------------------------------------------------------------------
> > --
> > > > > --- I'm using MET V4.1 and tried to specify use of a
forecast
> > > > > grid
> > > > > (GRIB1
> > > > > format) to define a verification masking region per the
> > > > > guidance on the User's Guide on p. 4-12 ("...any gridded
data
> > > > > file that MET can
> > > > read..."
> > > > > .
> > > > > I
> > > > > used the following entry in the config file:
> > > > >
> > > > > mask = {
> > > > > grid =
> > > > >
> > > > >
> >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012
> > -0
> > > > > 2-07_1
> > > > > 2:00:00"];
> > > > > poly = [];
> > > > > sid = [];
> > > > > };
> > > > >
> > > > > The ERROR dialog is as follows:
> > > > >
> > > > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > > > >
> > > > >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
> > 02
> > > > > -07_12
> > > > > :00:00"
> > > > > ERROR :
> > > > > ERROR : parse_grid_mask() -> the mask_grid requested
> > > > >
> > > > >
> >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
> > 02
> > > > > -07_12
> > > > > :00:00" is not defined.
> > > > > ERROR :
> > > > >
> > > > > The User's Guide mentions that the user must specify a
> > > > > description
> of
> > > > > the field to be used from the input file, but I wasn't sure
> > > > > how
> that
> > > > > info is input.
> > > > >
> > > > > Do you have any suggestions I can try to resolve this issue?
> > > > >
> > > > > Thanks.
> > > > >
> > > > > John Raby
> > > > > US Army Research Laboratory
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > Classification: UNCLASSIFIED
> > > > > Caveats: NONE
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > Classification: UNCLASSIFIED
> > > > Caveats: NONE
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
> >
> >
>
>
>
>
Classification: UNCLASSIFIED
Caveats: NONE
------------------------------------------------
Subject: MET Point-Stat mask_grid issue
From: John Halley Gotway
Time: Fri Mar 13 15:53:16 2015
John,
Great, I'm glad that all worked out! What a nice way to end the week.
I'll go ahead and resolve this ticket.
Thanks,
John
On Fri, Mar 13, 2015 at 3:49 PM, Raby, John W USA CIV via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
> John -
>
> I just finished running Point-Stat with three input forecast files
each
> with a
> different resolution (grid spacing), but for domains which are all
> centered on
> the same point and are all nested. I had generated the gen_poly_mask
NetCDF
> masking files using the forecast file for the innermost nest to
define the
> masking region. The config files for each of the three Point-Stat
runs
> specified the unique masking file which had been generated with that
> particular forecast file and it worked fine. I addition, I checked
the
> total
> number of matched pairs in the three Point-Stat CNT output files and
they
> were
> identical! This is what I was trying to achieve. The VX_MASK col had
the
> expected names. In IDV, the graphical depiction of the masking
region was
> as
> expected and all three regions were coincident.
>
> This will also me to compare the error stats for the forecasts
generated at
> three different grid spacings to evaluate the value added of higher
> resolution
> forecasts.
>
> Thanks for all your help.
>
> R/
> John
>
>
> -----Original Message-----
> From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> Sent: Friday, March 13, 2015 1:37 PM
> To: Raby, John W CIV USARMY ARL (US)
> Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> UPDATED
> info (UNCLASSIFIED)
>
> John,
>
> When you run Point-Stat, it reads the grid information from the
input
> forecast
> file. That grid definition includes the number of x/y points,
lat/lon
> origin,
> and several other parameters. When it reads the gridded masking
file, it
> also
> extracts the grid information from it. If those 2 grid definitions
differ
> in
> any way, Point-Stat will error out. So you need to run
gen_poly_mask for
> each
> forecast grid you want to evaluate.
>
> Thanks,
> John
>
> On Fri, Mar 13, 2015 at 1:17 PM, Raby, John W USA CIV via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> >
> > John -
> > Thanks. that correction fixed it.
> > Just ran:
> >
> > gen_poly_mask
> > Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ LAX_Dumais_D1_d03_poly.nc
\
> > -name DumaisD3mask
> >
> > With IDV, I visualize the file: LAX_Dumais_D2_d03_poly.nc and the
> > title of the graphic is DumaisD3mask
> >
> > Can this same .nc file be used as input to Point-Stat where I've
used
> > other input forecast grids as long as the forecast grid contains
the
> > area defined in the .nc file?
> >
> > R/
> > John
> > ________________________________________
> > From: John Halley Gotway via RT [met_help at ucar.edu]
> > Sent: Friday, March 13, 2015 1:00 PM
> > To: Raby, John W CIV USARMY ARL (US)
> > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
more
> > UPDATED info (UNCLASSIFIED)
> >
> > John,
> >
> > Just take a closer look at the usage statement for
gen_poly_mask...
> > you need to pass in 3 file names: input data_file name, input
> > mask_file name, and the output netcdf_file name.
> >
> > Try this...
> >
> > gen_poly_mask \
> > Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ LAX_Dumais_D3.nc \ -name
> > LAX_MASK
> >
> > Or use whatever value for the "-name" option you'd like.
> >
> > Thanks,
> > John
> >
> > On Fri, Mar 13, 2015 at 12:55 PM, Raby, John W USA CIV via RT <
> > met_help at ucar.edu> wrote:
> >
> > >
> > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > >
> > > John -
> > >
> > > I ran the following command:
> > >
> > > gen_poly_mask
> > Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ -name LAX_Dumais_D3
> > >
> > > but, no output is generated and all that gets printed to the
screen
> > > is
> > the
> > > usage.
> > >
> > > I then ran:
> > > gen_poly_mask
> > Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ -name LAX_Dumais_D3 -v
3
> > >
> > > To see if the verbosity would identify the problem, but again no
> > > output was generated and only the usage was printed to the
screen.
> > >
> > > I then tried:
> > >
> > > gen_poly_mask
> > Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00 \
> > > 131127A_WRFPRS_d03_2012-02-07_12:00:00 \ -name LAX_Dumais_D3.nc
V3
> > >
> > > Any explanatio for what is happening here?
> > >
> > > This time output was generated with name "V3" and when I looked
at
> > > V3 in IDV, the title of the graphic which shows my masking
region is
> > > "LAX_Dumais_D3". The graphic look correct. Can I use this same
mask
> > > file
> > in
> > > combination with other input forecast files as long as that area
is
> > > contained within the domain of the forecast files?
> > >
> > > R/
> > > John
> > > ________________________________________
> > > From: John Halley Gotway via RT [met_help at ucar.edu]
> > > Sent: Friday, March 13, 2015 12:26 PM
> > > To: Raby, John W CIV USARMY ARL (US)
> > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid issue
> > > more UPDATED info (UNCLASSIFIED)
> > >
> > > John,
> > >
> > > When you pass a lat/lon polyline file to gen_poly_mask, the
first
> > > line of that polyline file contains a "name" for the masking
region.
> > > That name
> > is
> > > used as the variable name in the NetCDF output file. However,
when
> > > you
> > use
> > > a gridded data file to define the masking region, we have no
good
> > > default value for what name you'd like to assign to that mask.
That
> > > error
> > message
> > > is instructing you to use the "-name" command line switch to
define it.
> > > You might consider "-name NEST" or perhaps "-name D03_Grid".
> > > Something like that.
> > >
> > > When you use the gen_poly_mask output in Point-Stat, it'll
extract
> > > that name from the mask file and write it to the VX_MASK column
in
> > > the Point-Stat output files. So just choose whatever name you'd
> > > like to
> > appear
> > > in the Point-Stat output files.
> > >
> > > Make sense?
> > >
> > > Thanks,
> > > John
> > >
> > > On Fri, Mar 13, 2015 at 12:17 PM, Raby, John W USA CIV via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > > >
> > > > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078
>
> > > >
> > > > John -
> > > >
> > > > Just ran the following command:
> > > >
> > > > gen_poly_mask
> > Dumais_FDDA_20120207/131127A_WRFPRS_d01_2012-02-07_12:00:00
> > > > \ 131127A_WRFPRS_d03_2012-02-07_12:00:00 \
> > > > LAX_Dumais_D1_d03_poly.nc
> > > >
> > > > Received an error as follows:
> > > >
> > > > ERROR :
> > > > ERROR : write_netcdf() -> the masking variable name must be
set
> > > > using
> > > the
> > > > "-name" command line argument.
> > > > ERROR :
> > > >
> > > > The output .nc file was generated. What is the masking
variable?
> > > > The
> > > usage
> > > > says that the -name is optional. Is the output OK?
> > > >
> > > > R/
> > > > John
> > > > ________________________________________
> > > > From: John Halley Gotway via RT [met_help at ucar.edu]
> > > > Sent: Friday, March 13, 2015 11:25 AM
> > > > To: Raby, John W CIV USARMY ARL (US)
> > > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> > > > more UPDATED info (UNCLASSIFIED)
> > > >
> > > > John,
> > > >
> > > > Thanks for letting me, and yes, please let me know what your
> > > > testing reveals.
> > > >
> > > > I apologize the for user's guide being out-of-date and/or
unclear.
> > We'll
> > > > be rewriting that section when we replace gen_poly_mask with
> > gen_vx_mask
> > > > anyway. But we do have larger issues with the maintenance of
our
> > user's
> > > > guide. We are hoping/planning to make changes in how we do
our
> > > > documentation to make it much easier to keep it in line with
the
> > > > code itself.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Fri, Mar 13, 2015 at 11:12 AM, Raby, John W USA CIV via RT
<
> > > > met_help at ucar.edu> wrote:
> > > >
> > > > >
> > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078 >
> > > > >
> > > > > Classification: UNCLASSIFIED
> > > > > Caveats: NONE
> > > > >
> > > > > John -
> > > > >
> > > > > Thanks. Your answers address all my questions very well.
> > > > >
> > > > > My lat/lon polyline has only four points so I concur that
the
> > > > > runtime
> > > is
> > > > a
> > > > > non-issue for sure.
> > > > >
> > > > > I am very interested in your suggestion about using the
> > > > > gen_poly_mask
> > > > tool
> > > > > and
> > > > > providing the GRIB file of the smaller forecast as the input
> > > > "mask_file". I
> > > > > just generated the usage in an xterm and see the wording you
> > > > > show
> > > below.
> > > > I
> > > > > should have checked there. Maybe an edit change for the
User's
> > > > > Guide
> > p.
> > > > > 3-24
> > > > > (V4.1 or later update) would also help.
> > > > >
> > > > > Regarding the different matched pair counts, I have the MPR
> > > > > files for
> > > > the 3
> > > > > runs and could do a "diff" among them and some other
exploring
> > > > > to
> > look
> > > > for
> > > > > the
> > > > > differences. I will let you know what differences I find. I
> > > > > suspect
> > > that
> > > > > the
> > > > > first possibility you mention below is the more likely
explanation.
> > If
> > > > the
> > > > > input forecast grid is referenced when converting the
polygon
> > > > > points
> > > into
> > > > > the
> > > > > x/y values and the same with the observation locations, then
> > > > > that x/y
> > > > grid
> > > > > space might differ some from that which would be referenced
when
> > > > > a
> > > > forecast
> > > > > grid with closer grid spacing is used with the same polygon
> > > > > points
> > and
> > > > > observations.
> > > > >
> > > > > I am hoping that by using the gen_poly_mask tool and
providing
> > > > > the
> > GRIB
> > > > > file
> > > > > of the smaller forecast as the input "mask_file", a more
> > > > > accurate replication of the desired masking region might
result
> > > > > which results in more consistency in the numbers of matched
> > > > > pairs. I will let you know the outcome.
> > > > >
> > > > > R/
> > > > > John
> > > > >
> > > > > -----Original Message-----
> > > > > From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
> > > > > Sent: Friday, March 13, 2015 10:30 AM
> > > > > To: Raby, John W CIV USARMY ARL (US)
> > > > > Subject: Re: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
issue
> > more
> > > > > UPDATED
> > > > > info (UNCLASSIFIED)
> > > > >
> > > > > John,
> > > > >
> > > > > Glad you were able to make some progress. I see a couple of
> > remaining
> > > > > questions in there...
> > > > >
> > > > > "Using the output of gen_poly_mask and the lat/lon polyline
> > > > > itself
> > > > produces
> > > > > the same result. Is that OK?"
> > > > >
> > > > > My answer is... that's great, and I'm relieved they do! The
> > > > > main
> > > reason
> > > > to
> > > > > use gen_poly_mask is to make the MET statistics tools (i.e.
> > point_stat
> > > or
> > > > > grid_stat) run a bit faster. It can be slow to figure out
which
> > points
> > > > are
> > > > > inside/outside a polyline region if the polyline contains
> > > > > thousands
> > of
> > > > > point.
> > > > > In that case, you'd run gen_poly_mask to create the mask
once
> > > > > rather
> > > than
> > > > > having to recreate it each time you run a MET statistics
tool.
> > > > > But if your polyline only contains a small number of points,
> > > > > running gen_poly_mask first would only save you a second or
two.
> > > > >
> > > > > Next, you were trying to use a smaller grid to create a mask
for
> > > > > a
> > > larger
> > > > > grid. The gen_poly_mask tool does support this. By the
way,
> > > > > since METv5.0, we've realized that gen-"poly"-mask is poorly
> > > > > named, and in the next release it'll be replaced by a more
> > > > > generalized tool named gen_vx_mask. It
> > > > include
> > > > > some more flexible options.
> > > > >
> > > > > But if you read the usage statement for gen_poly_mask,
the...
> > > > > "mask_file" is an ASCII Lat/Lon polyline file or gridded
data
> > > > > file defining the masking region
> > > > >
> > > > > So you can pass a gridded data file for the nest as the
"mask_file"
> > > > rather
> > > > > than having to create a set of lat/lon points. I'd suggest
> > > > > going
> > that
> > > > > route.
> > > > >
> > > > > Here's an example of that using data from the MET tarball:
> > > > > cd met-5.0
> > > > > bin/gen_poly_mask \
> > > > > data/sample_fcst/2005080700/wrfprs_ruc13_12.tm00_G212 \
> > > > > data/sample_fcst/2009123112/arw-fer-
gep1/d01_2009123112_02400.gr
> > > > > ib \ mask.nc -name MY_MASK
> > > > >
> > > > > Unfortunately, I don't think I fully understand your last
> > > > > question as
> > > to
> > > > > why
> > > > > you're getting slightly different matched pair counts (621,
619
> > > > > and
> > > 617).
> > > > > But
> > > > > my guess is one of two possibilities. Either...
> > > > > (1) The polyline mask for the different grids is producing
> > > > > slightly different results and the differences lie along the
> > > > > boundary of the mask.
> > > > > (2) There's some missing data values somewhere in the
forecast
> > > > > and observations causing slightly different matched pairs.
> > > > >
> > > > > I'd suggest configuring point_stat to dump out the MPR lines
for
> > those
> > > > > three
> > > > > runs. Then we could take a close look at them to see where
the
> > > > differences
> > > > > lie. Identifying the stations where the differences occur
is
> > > > > the
> > first
> > > > > step
> > > > > in finding an explanation.
> > > > >
> > > > > Thanks,
> > > > > John
> > > > >
> > > > >
> > > > >
> > > > > On Fri, Mar 13, 2015 at 8:02 AM, Raby, John W USA CIV via RT
<
> > > > > met_help at ucar.edu> wrote:
> > > > >
> > > > > >
> > > > > > <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=71078
> > > > > > >
> > > > > >
> > > > > > Classification: UNCLASSIFIED
> > > > > > Caveats: NONE
> > > > > >
> > > > > > I tried another method to define the masking region using
the
> > > > > > gen_poly_mask tool using my input forecast file and a poly
> > > > > > file
> > which
> > > > > > defines the rectangular subdomain in the center of the two
> > > > > > larger input forecast domains and was successful, but the
> > > > > > results I get
> > are
> > > > > > no different than when I use the poly text file itself.
This
> > > > > > brings
> > > up
> > > > > > another question about the results I get with the poly
text file.
> > If
> > > I
> > > > > > run Point-Stat with three different input forecast grids
(all
> > > > > > generated from the same WRF run over a set of nested
domains
> > > > > > which have a common center) with three different
resolutions
> > > > > > (grid
> > spacing)
> > > > > > and specify the same masking region using the text poly
file
> > > > > > which defines the domain of the innermost nest, the three
CNT
> > > > > > output
> > files
> > > > > > show three different values (621, 619 and 617) for the
TOTAL
> > > > > > number
> > > of
> > > > > > matched pairs for the variable TMP/Z2. Why is this the
case
> > > > > > when
> > the
> > > > > > scoring is being done over the same masking region? My
> > > > > > expectation
> > > was
> > > > > > that the TOTAL number of matched pairs should all be the
same.
> > > > > >
> > > > > > Thanks.
> > > > > >
> > > > > > R/
> > > > > > John
> > > > > >
> > > > > > -----Original Message-----
> > > > > > From: Raby, John W CIV USARMY ARL (US)
> > > > > > Sent: Thursday, March 12, 2015 1:17 PM
> > > > > > To: met_help at ucar.edu
> > > > > > Subject: RE: [rt.rap.ucar.edu #71078] AutoReply: mask_grid
> > > > > > issue UPDATED info
> > > > > >
> > > > > > I was able to locate some info in the README file in the
> > > > > > software distribution folder which describes the entries
in
> > > > > > the config file
> > to
> > > > > > describe the field in the forecast file to be used. I
modified
> > > > > > my config file as follows:
> > > > > >
> > > > > > mask = {
> > > > > > grid = [];
> > > > > > poly =
> > > > > >
> > > > > >
> > >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012
> > > -0
> > > > > > 2-07_1
> > > > > > 2:00:00 {name = \"TMP\"; level = \"Z2\";}"
> > > > > > ];
> > > > > > sid = [];
> > > > > > };
> > > > > >
> > > > > > Point-Stat executed for a longer period, but no output.
Had an
> > ERROR
> > > > > > as
> > > > > > follows:
> > > > > >
> > > > > > ERROR :
> > > > > > ERROR : parse_poly_mask() -> the masking grid does not
match
> > > > > > the input grid.
> > > > > > ERROR :
> > > > > >
> > > > > > This appears to present an issue with the way I want to
use
> > > > > > the
> > > > > mask_grid.
> > > > > > My intent was to score the input forecast file, but only
over
> > > > > > a smaller domain centered inside the domain of the input
> > > > > > forecast
> > file.
> > > > > > Any thoughts on this development and how I can proceed?
> > > > > >
> > > > > > Thanks.
> > > > > > John
> > > > > > _______________________________________
> > > > > > From: met_help at ucar.edu via RT [met_help at ucar.edu]
> > > > > > Sent: Thursday, March 12, 2015 11:44 AM
> > > > > > To: Raby, John W CIV USARMY ARL (US)
> > > > > > Subject: [rt.rap.ucar.edu #71078] AutoReply: MET Point-
Stat
> > > mask_grid
> > > > > > issue
> > > > > >
> > > > > > Greetings,
> > > > > >
> > > > > > This message has been automatically generated in response
to
> > > > > > the creation of a trouble ticket regarding:
> > > > > > "MET Point-Stat mask_grid issue", a summary of
which
> > appears
> > > > > below.
> > > > > >
> > > > > > There is no need to reply to this message right now. Your
> > > > > > ticket
> > has
> > > > > > been assigned an ID of [rt.rap.ucar.edu #71078].
> > > > > >
> > > > > > Please include the string:
> > > > > >
> > > > > > [rt.rap.ucar.edu #71078]
> > > > > >
> > > > > > in the subject line of all future correspondence about
this
> issue.
> > To
> > > > > > do so, you may reply to this message.
> > > > > >
> > > > > > Thank you,
> > > > > > met_help at ucar.edu
> > > > > >
> > > > > >
> > >
--------------------------------------------------------------------
> > > --
> > > > > > --- I'm using MET V4.1 and tried to specify use of a
forecast
> > > > > > grid
> > > > > > (GRIB1
> > > > > > format) to define a verification masking region per the
> > > > > > guidance on the User's Guide on p. 4-12 ("...any gridded
data
> > > > > > file that MET can
> > > > > read..."
> > > > > > .
> > > > > > I
> > > > > > used the following entry in the config file:
> > > > > >
> > > > > > mask = {
> > > > > > grid =
> > > > > >
> > > > > >
> > >
["/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012
> > > -0
> > > > > > 2-07_1
> > > > > > 2:00:00"];
> > > > > > poly = [];
> > > > > > sid = [];
> > > > > > };
> > > > > >
> > > > > > The ERROR dialog is as follows:
> > > > > >
> > > > > > DEBUG 4: parse_grid_mask() -> parsing grid mask
> > > > > >
> > > > > >
> > >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
> > > 02
> > > > > > -07_12
> > > > > > :00:00"
> > > > > > ERROR :
> > > > > > ERROR : parse_grid_mask() -> the mask_grid requested
> > > > > >
> > > > > >
> > >
"/home/jraby/MET_WRFpostprd/20120207_D_FDDA/131127A_WRFPRS_d03_2012-
> > > 02
> > > > > > -07_12
> > > > > > :00:00" is not defined.
> > > > > > ERROR :
> > > > > >
> > > > > > The User's Guide mentions that the user must specify a
> > > > > > description
> > of
> > > > > > the field to be used from the input file, but I wasn't
sure
> > > > > > how
> > that
> > > > > > info is input.
> > > > > >
> > > > > > Do you have any suggestions I can try to resolve this
issue?
> > > > > >
> > > > > > Thanks.
> > > > > >
> > > > > > John Raby
> > > > > > US Army Research Laboratory
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > Classification: UNCLASSIFIED
> > > > > > Caveats: NONE
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > Classification: UNCLASSIFIED
> > > > > Caveats: NONE
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
> >
> >
> >
>
>
> Classification: UNCLASSIFIED
> Caveats: NONE
>
>
>
>
------------------------------------------------
More information about the Met_help
mailing list